cephi/ampere_research
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

as-caida data

Browse Source
This commit is contained in:
cephi 2024-12-17 23:29:19 -05:00
parent e22addba44
commit 65e55bd251
865 changed files with 18611 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
pytorch/batch.py
View File

@ -117,7 +117,7 @@ elif args.matrix_type == MatrixType.SYNTHETIC:
parameter_list = enumerate([(size, density)
for size in args.synthetic_size
for density in args.synthetic_density
if size ** 2 * density <= 30000000])
if size ** 2 * density <= 50000000])
#for i, matrix in enumerate(glob.glob(f'{args.matrix_dir.rstrip("/")}/*.mtx')):
for i, parameter in parameter_list:

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_005.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5806, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 10.171998977661133, "TIME_S_1KI": 1.7519805335275807, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 312.83616289138786, "W": 23.626422472883572, "J_1KI": 53.88152995029071, "W_1KI": 4.069311483445328, "W_D": 5.210422472883575, "J_D": 68.9909179153442, "W_D_1KI": 0.8974203363561101, "J_D_1KI": 0.15456774653050467}

65
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_005.output Normal file
View File

@ -0,0 +1,65 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_005.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 0.1808183193206787}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.7415, 0.8054, 0.6431, ..., 0.7043, 0.2095, 0.2852])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 0.1808183193206787 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5806 -m matrices/as-caida_pruned/as-caida_G_005.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 10.171998977661133}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.0310, 0.3080, 0.7594, ..., 0.0941, 0.5225, 0.9795])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 10.171998977661133 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.0310, 0.3080, 0.7594, ..., 0.0941, 0.5225, 0.9795])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 10.171998977661133 seconds
[20.56, 20.64, 20.44, 20.48, 20.48, 20.44, 20.6, 20.64, 20.64, 20.76]
[20.8, 20.88, 21.04, 25.72, 26.8, 29.92, 30.76, 28.4, 26.88, 24.6, 24.56, 24.56, 24.72]
13.240945100784302
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5806, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 10.171998977661133, 'TIME_S_1KI': 1.7519805335275807, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 312.83616289138786, 'W': 23.626422472883572}
[20.56, 20.64, 20.44, 20.48, 20.48, 20.44, 20.6, 20.64, 20.64, 20.76, 20.2, 20.36, 20.24, 20.2, 20.24, 20.36, 20.4, 20.56, 20.6, 20.48]
368.31999999999994
18.415999999999997
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5806, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 10.171998977661133, 'TIME_S_1KI': 1.7519805335275807, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 312.83616289138786, 'W': 23.626422472883572, 'J_1KI': 53.88152995029071, 'W_1KI': 4.069311483445328, 'W_D': 5.210422472883575, 'J_D': 68.9909179153442, 'W_D_1KI': 0.8974203363561101, 'J_D_1KI': 0.15456774653050467}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_010.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5720, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 10.46983003616333, "TIME_S_1KI": 1.8303898664621205, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 327.3828331851959, "W": 23.035559333910587, "J_1KI": 57.23476104636292, "W_1KI": 4.027195687746606, "W_D": 4.593559333910587, "J_D": 65.28395717859267, "W_D_1KI": 0.803069813620732, "J_D_1KI": 0.14039682056306504}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_010.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_010.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 0.23993563652038574}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.3498, 0.5888, 0.0645, ..., 0.9730, 0.2168, 0.4693])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 0.23993563652038574 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4376 -m matrices/as-caida_pruned/as-caida_G_010.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 8.03219199180603}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.9782, 0.4472, 0.6352, ..., 0.3530, 0.3134, 0.8485])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 8.03219199180603 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5720 -m matrices/as-caida_pruned/as-caida_G_010.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 10.46983003616333}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.5450, 0.3229, 0.9483, ..., 0.0104, 0.7843, 0.5252])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 10.46983003616333 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.5450, 0.3229, 0.9483, ..., 0.0104, 0.7843, 0.5252])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 10.46983003616333 seconds
[20.48, 20.36, 20.36, 20.28, 20.44, 20.76, 20.8, 20.72, 20.64, 20.24]
[20.2, 20.04, 22.12, 23.48, 25.88, 25.88, 27.08, 28.04, 26.28, 26.48, 25.0, 24.88, 24.72, 24.44]
14.212063550949097
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5720, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_010', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 74994, 'MATRIX_DENSITY': 7.616375021864427e-05, 'TIME_S': 10.46983003616333, 'TIME_S_1KI': 1.8303898664621205, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 327.3828331851959, 'W': 23.035559333910587}
[20.48, 20.36, 20.36, 20.28, 20.44, 20.76, 20.8, 20.72, 20.64, 20.24, 20.28, 20.36, 20.52, 20.68, 20.48, 20.48, 20.52, 20.4, 20.32, 20.44]
368.84000000000003
18.442
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5720, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_010', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 74994, 'MATRIX_DENSITY': 7.616375021864427e-05, 'TIME_S': 10.46983003616333, 'TIME_S_1KI': 1.8303898664621205, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 327.3828331851959, 'W': 23.035559333910587, 'J_1KI': 57.23476104636292, 'W_1KI': 4.027195687746606, 'W_D': 4.593559333910587, 'J_D': 65.28395717859267, 'W_D_1KI': 0.803069813620732, 'J_D_1KI': 0.14039682056306504}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_015.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5449, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 10.42112112045288, "TIME_S_1KI": 1.9124832300335624, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 326.5276795768738, "W": 22.970903604955094, "J_1KI": 59.924330992269006, "W_1KI": 4.21561820608462, "W_D": 4.577903604955093, "J_D": 65.0741593434811, "W_D_1KI": 0.8401364663158548, "J_D_1KI": 0.15418177029103594}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_015.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_015.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 0.24444031715393066}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.0041, 0.0059, 0.8299, ..., 0.3077, 0.8545, 0.8513])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 0.24444031715393066 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4295 -m matrices/as-caida_pruned/as-caida_G_015.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 8.275424003601074}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.5048, 0.0128, 0.9259, ..., 0.5690, 0.7343, 0.9731])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 8.275424003601074 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5449 -m matrices/as-caida_pruned/as-caida_G_015.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 10.42112112045288}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.2882, 0.9342, 0.0969, ..., 0.6573, 0.4161, 0.3369])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 10.42112112045288 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.2882, 0.9342, 0.0969, ..., 0.6573, 0.4161, 0.3369])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 10.42112112045288 seconds
[20.72, 20.48, 20.48, 20.32, 20.36, 20.36, 20.28, 20.36, 20.36, 20.24]
[20.4, 20.36, 20.28, 25.0, 25.96, 28.32, 29.16, 26.92, 25.8, 24.32, 24.4, 24.24, 24.24, 24.24]
14.21483826637268
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5449, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_015', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 77124, 'MATRIX_DENSITY': 7.832697378273889e-05, 'TIME_S': 10.42112112045288, 'TIME_S_1KI': 1.9124832300335624, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 326.5276795768738, 'W': 22.970903604955094}
[20.72, 20.48, 20.48, 20.32, 20.36, 20.36, 20.28, 20.36, 20.36, 20.24, 20.48, 20.36, 20.32, 20.2, 20.16, 20.6, 20.76, 20.72, 20.8, 20.44]
367.86
18.393
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5449, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_015', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 77124, 'MATRIX_DENSITY': 7.832697378273889e-05, 'TIME_S': 10.42112112045288, 'TIME_S_1KI': 1.9124832300335624, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 326.5276795768738, 'W': 22.970903604955094, 'J_1KI': 59.924330992269006, 'W_1KI': 4.21561820608462, 'W_D': 4.577903604955093, 'J_D': 65.0741593434811, 'W_D_1KI': 0.8401364663158548, 'J_D_1KI': 0.15418177029103594}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_020.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5290, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 10.617355108261108, "TIME_S_1KI": 2.0070614571382057, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 322.70390956878657, "W": 22.650037252323052, "J_1KI": 61.0026294080882, "W_1KI": 4.281670558095095, "W_D": 4.262037252323051, "J_D": 60.72290604782095, "W_D_1KI": 0.8056781195317676, "J_D_1KI": 0.15230210199088234}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_020.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_020.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 0.2548098564147949}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.8564, 0.5356, 0.2431, ..., 0.6892, 0.8837, 0.8255])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 0.2548098564147949 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4120 -m matrices/as-caida_pruned/as-caida_G_020.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 8.176596403121948}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.6319, 0.0708, 0.0554, ..., 0.3150, 0.6632, 0.2317])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 8.176596403121948 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5290 -m matrices/as-caida_pruned/as-caida_G_020.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 10.617355108261108}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.4957, 0.1363, 0.6360, ..., 0.3901, 0.1824, 0.9506])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 10.617355108261108 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.4957, 0.1363, 0.6360, ..., 0.3901, 0.1824, 0.9506])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 10.617355108261108 seconds
[20.12, 20.28, 20.16, 20.16, 20.12, 20.24, 20.32, 20.4, 20.4, 20.48]
[20.6, 20.72, 20.8, 22.2, 23.96, 25.28, 26.36, 26.92, 26.56, 25.12, 25.08, 25.24, 25.24, 25.32]
14.247389793395996
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5290, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_020', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 80948, 'MATRIX_DENSITY': 8.221062021893506e-05, 'TIME_S': 10.617355108261108, 'TIME_S_1KI': 2.0070614571382057, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 322.70390956878657, 'W': 22.650037252323052}
[20.12, 20.28, 20.16, 20.16, 20.12, 20.24, 20.32, 20.4, 20.4, 20.48, 20.32, 20.36, 20.48, 20.48, 20.68, 20.76, 20.64, 20.68, 20.8, 20.68]
367.76000000000005
18.388
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5290, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_020', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 80948, 'MATRIX_DENSITY': 8.221062021893506e-05, 'TIME_S': 10.617355108261108, 'TIME_S_1KI': 2.0070614571382057, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 322.70390956878657, 'W': 22.650037252323052, 'J_1KI': 61.0026294080882, 'W_1KI': 4.281670558095095, 'W_D': 4.262037252323051, 'J_D': 60.72290604782095, 'W_D_1KI': 0.8056781195317676, 'J_D_1KI': 0.15230210199088234}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_025.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5009, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 10.436067342758179, "TIME_S_1KI": 2.0834632347291233, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 339.9102258396148, "W": 23.870320724339745, "J_1KI": 67.85989735268812, "W_1KI": 4.7654862695827, "W_D": 5.56032072433975, "J_D": 79.17823539018632, "W_D_1KI": 1.110066026021112, "J_D_1KI": 0.22161429946518504}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_025.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_025.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 0.2360525131225586}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.8761, 0.0368, 0.8631, ..., 0.6340, 0.9685, 0.1396])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 0.2360525131225586 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4448 -m matrices/as-caida_pruned/as-caida_G_025.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 9.323699712753296}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.4095, 0.9128, 0.5370, ..., 0.1298, 0.9549, 0.0765])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 9.323699712753296 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5009 -m matrices/as-caida_pruned/as-caida_G_025.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 10.436067342758179}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.6896, 0.4674, 0.9391, ..., 0.8690, 0.1471, 0.0542])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 10.436067342758179 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.6896, 0.4674, 0.9391, ..., 0.8690, 0.1471, 0.0542])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 10.436067342758179 seconds
[20.2, 20.32, 20.52, 20.52, 20.64, 20.64, 20.64, 20.64, 20.64, 20.52]
[20.44, 20.72, 20.72, 23.84, 26.4, 29.12, 30.32, 31.08, 27.36, 26.24, 24.52, 25.04, 25.36, 25.64]
14.239868402481079
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5009, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_025', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 85850, 'MATRIX_DENSITY': 8.718908121010495e-05, 'TIME_S': 10.436067342758179, 'TIME_S_1KI': 2.0834632347291233, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 339.9102258396148, 'W': 23.870320724339745}
[20.2, 20.32, 20.52, 20.52, 20.64, 20.64, 20.64, 20.64, 20.64, 20.52, 20.2, 20.16, 20.16, 20.16, 20.2, 20.24, 20.16, 20.2, 19.92, 19.96]
366.19999999999993
18.309999999999995
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5009, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_025', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 85850, 'MATRIX_DENSITY': 8.718908121010495e-05, 'TIME_S': 10.436067342758179, 'TIME_S_1KI': 2.0834632347291233, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 339.9102258396148, 'W': 23.870320724339745, 'J_1KI': 67.85989735268812, 'W_1KI': 4.7654862695827, 'W_D': 5.56032072433975, 'J_D': 79.17823539018632, 'W_D_1KI': 1.110066026021112, 'J_D_1KI': 0.22161429946518504}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_030.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4924, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 10.37181830406189, "TIME_S_1KI": 2.106380646641326, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 321.23179523468013, "W": 22.538047273928385, "J_1KI": 65.23797628649068, "W_1KI": 4.577182630773433, "W_D": 4.022047273928383, "J_D": 57.325705755233706, "W_D_1KI": 0.8168251977921167, "J_D_1KI": 0.16588651458003995}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_030.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_030.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 0.27124905586242676}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.5482, 0.1173, 0.8258, ..., 0.9814, 0.8155, 0.1910])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 0.27124905586242676 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3870 -m matrices/as-caida_pruned/as-caida_G_030.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 8.251606702804565}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.8510, 0.7355, 0.4606, ..., 0.8447, 0.7677, 0.2060])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 8.251606702804565 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4924 -m matrices/as-caida_pruned/as-caida_G_030.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 10.37181830406189}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.8967, 0.1733, 0.8053, ..., 0.0536, 0.8153, 0.7243])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 10.37181830406189 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.8967, 0.1733, 0.8053, ..., 0.0536, 0.8153, 0.7243])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 10.37181830406189 seconds
[20.32, 20.36, 20.4, 20.36, 20.52, 20.64, 20.56, 20.6, 20.48, 20.4]
[20.12, 20.28, 21.16, 22.48, 22.48, 24.72, 25.72, 26.68, 26.52, 26.0, 25.68, 25.52, 25.08, 25.04]
14.252867221832275
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4924, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_030', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 86850, 'MATRIX_DENSITY': 8.820467912752026e-05, 'TIME_S': 10.37181830406189, 'TIME_S_1KI': 2.106380646641326, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 321.23179523468013, 'W': 22.538047273928385}
[20.32, 20.36, 20.4, 20.36, 20.52, 20.64, 20.56, 20.6, 20.48, 20.4, 20.56, 20.4, 20.44, 20.64, 20.64, 20.56, 20.84, 20.76, 20.92, 21.12]
370.32000000000005
18.516000000000002
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4924, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_030', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 86850, 'MATRIX_DENSITY': 8.820467912752026e-05, 'TIME_S': 10.37181830406189, 'TIME_S_1KI': 2.106380646641326, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 321.23179523468013, 'W': 22.538047273928385, 'J_1KI': 65.23797628649068, 'W_1KI': 4.577182630773433, 'W_D': 4.022047273928383, 'J_D': 57.325705755233706, 'W_D_1KI': 0.8168251977921167, 'J_D_1KI': 0.16588651458003995}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_035.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4906, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 10.369710445404053, "TIME_S_1KI": 2.113679259152885, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 320.3935369491577, "W": 22.512475730235575, "J_1KI": 65.30646900716626, "W_1KI": 4.5887639075082705, "W_D": 4.301475730235577, "J_D": 61.217834938526146, "W_D_1KI": 0.8767785834153234, "J_D_1KI": 0.17871556938755065}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_035.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_035.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 0.28984951972961426}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.2495, 0.4746, 0.6508, ..., 0.6030, 0.3808, 0.6963])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 0.28984951972961426 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3622 -m matrices/as-caida_pruned/as-caida_G_035.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 7.750367879867554}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.1862, 0.1434, 0.8787, ..., 0.7704, 0.8925, 0.8878])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 7.750367879867554 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4906 -m matrices/as-caida_pruned/as-caida_G_035.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 10.369710445404053}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.7972, 0.7657, 0.0835, ..., 0.8008, 0.2416, 0.9619])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 10.369710445404053 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.7972, 0.7657, 0.0835, ..., 0.8008, 0.2416, 0.9619])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 10.369710445404053 seconds
[20.08, 19.84, 19.84, 19.92, 19.96, 20.04, 20.08, 20.2, 20.2, 20.2]
[20.36, 20.32, 21.32, 22.76, 24.92, 24.92, 25.72, 26.36, 26.24, 25.64, 24.64, 24.68, 24.72, 24.56]
14.231821537017822
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4906, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_035', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 87560, 'MATRIX_DENSITY': 8.892575364888514e-05, 'TIME_S': 10.369710445404053, 'TIME_S_1KI': 2.113679259152885, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 320.3935369491577, 'W': 22.512475730235575}
[20.08, 19.84, 19.84, 19.92, 19.96, 20.04, 20.08, 20.2, 20.2, 20.2, 20.36, 20.4, 20.56, 20.4, 20.6, 20.6, 20.6, 20.28, 20.12, 20.52]
364.21999999999997
18.211
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4906, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_035', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 87560, 'MATRIX_DENSITY': 8.892575364888514e-05, 'TIME_S': 10.369710445404053, 'TIME_S_1KI': 2.113679259152885, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 320.3935369491577, 'W': 22.512475730235575, 'J_1KI': 65.30646900716626, 'W_1KI': 4.5887639075082705, 'W_D': 4.301475730235577, 'J_D': 61.217834938526146, 'W_D_1KI': 0.8767785834153234, 'J_D_1KI': 0.17871556938755065}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_040.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4720, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 10.206109046936035, "TIME_S_1KI": 2.1623112387576344, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 318.96216671943665, "W": 22.387743007233112, "J_1KI": 67.57673023716879, "W_1KI": 4.743165891362947, "W_D": 4.006743007233112, "J_D": 57.084782090902344, "W_D_1KI": 0.8488862303459983, "J_D_1KI": 0.1798487776156776}

65
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_040.output Normal file
View File

@ -0,0 +1,65 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_040.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 0.222426176071167}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.9066, 0.3947, 0.5988, ..., 0.8238, 0.0350, 0.4398])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 0.222426176071167 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4720 -m matrices/as-caida_pruned/as-caida_G_040.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 10.206109046936035}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.1500, 0.4836, 0.1224, ..., 0.2519, 0.6750, 0.1609])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 10.206109046936035 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.1500, 0.4836, 0.1224, ..., 0.2519, 0.6750, 0.1609])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 10.206109046936035 seconds
[20.2, 20.0, 20.04, 20.24, 20.28, 20.28, 20.56, 20.56, 20.4, 20.4]
[20.32, 20.48, 20.84, 22.0, 24.12, 26.52, 27.16, 27.04, 25.84, 24.28, 24.16, 24.2, 24.2, 24.04]
14.247178316116333
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4720, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_040', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89658, 'MATRIX_DENSITY': 9.105647807962247e-05, 'TIME_S': 10.206109046936035, 'TIME_S_1KI': 2.1623112387576344, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 318.96216671943665, 'W': 22.387743007233112}
[20.2, 20.0, 20.04, 20.24, 20.28, 20.28, 20.56, 20.56, 20.4, 20.4, 20.48, 20.48, 20.52, 20.6, 20.76, 20.68, 20.56, 20.64, 20.4, 20.16]
367.62
18.381
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4720, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_040', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89658, 'MATRIX_DENSITY': 9.105647807962247e-05, 'TIME_S': 10.206109046936035, 'TIME_S_1KI': 2.1623112387576344, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 318.96216671943665, 'W': 22.387743007233112, 'J_1KI': 67.57673023716879, 'W_1KI': 4.743165891362947, 'W_D': 4.006743007233112, 'J_D': 57.084782090902344, 'W_D_1KI': 0.8488862303459983, 'J_D_1KI': 0.1798487776156776}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_045.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4905, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 10.552767515182495, "TIME_S_1KI": 2.1514306860718646, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 328.58112309455873, "W": 23.14675123767438, "J_1KI": 66.9890159214187, "W_1KI": 4.719011465377039, "W_D": 4.7337512376743796, "J_D": 67.19825526070599, "W_D_1KI": 0.9650868986084362, "J_D_1KI": 0.196755738758091}

105
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_045.output Normal file
View File

@ -0,0 +1,105 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_045.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 0.2738626003265381}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.8777, 0.5124, 0.0822, ..., 0.9706, 0.3708, 0.5874])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 0.2738626003265381 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3834 -m matrices/as-caida_pruned/as-caida_G_045.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 8.640145778656006}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.9169, 0.5590, 0.9513, ..., 0.6480, 0.9706, 0.8048])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 8.640145778656006 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4659 -m matrices/as-caida_pruned/as-caida_G_045.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 9.972679376602173}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.6180, 0.9542, 0.9412, ..., 0.9357, 0.2218, 0.1163])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 9.972679376602173 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4905 -m matrices/as-caida_pruned/as-caida_G_045.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 10.552767515182495}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.7047, 0.9313, 0.0358, ..., 0.9576, 0.8194, 0.2072])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 10.552767515182495 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.7047, 0.9313, 0.0358, ..., 0.9576, 0.8194, 0.2072])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 10.552767515182495 seconds
[20.24, 20.28, 20.32, 20.28, 20.64, 20.72, 20.8, 20.96, 20.88, 20.6]
[20.6, 20.52, 20.28, 22.6, 24.08, 26.32, 27.4, 28.32, 27.04, 26.4, 25.72, 25.72, 25.88, 25.88]
14.195561170578003
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4905, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_045', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89152, 'MATRIX_DENSITY': 9.054258553341032e-05, 'TIME_S': 10.552767515182495, 'TIME_S_1KI': 2.1514306860718646, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 328.58112309455873, 'W': 23.14675123767438}
[20.24, 20.28, 20.32, 20.28, 20.64, 20.72, 20.8, 20.96, 20.88, 20.6, 20.8, 20.4, 20.12, 19.96, 20.16, 20.24, 20.4, 20.6, 20.48, 20.4]
368.26
18.413
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4905, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_045', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89152, 'MATRIX_DENSITY': 9.054258553341032e-05, 'TIME_S': 10.552767515182495, 'TIME_S_1KI': 2.1514306860718646, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 328.58112309455873, 'W': 23.14675123767438, 'J_1KI': 66.9890159214187, 'W_1KI': 4.719011465377039, 'W_D': 4.7337512376743796, 'J_D': 67.19825526070599, 'W_D_1KI': 0.9650868986084362, 'J_D_1KI': 0.196755738758091}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_050.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4641, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 10.064989805221558, "TIME_S_1KI": 2.16871144262477, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 338.8994520378113, "W": 23.81807615970891, "J_1KI": 73.02293730614335, "W_1KI": 5.132100012865527, "W_D": 5.377076159708913, "J_D": 76.5086210939885, "W_D_1KI": 1.1586029217213776, "J_D_1KI": 0.24964510271953838}

65
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_050.output Normal file
View File

@ -0,0 +1,65 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_050.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 0.22620630264282227}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.0575, 0.9187, 0.6279, ..., 0.6013, 0.5410, 0.6623])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 0.22620630264282227 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4641 -m matrices/as-caida_pruned/as-caida_G_050.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 10.064989805221558}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.7076, 0.1289, 0.6333, ..., 0.8099, 0.6262, 0.3286])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 10.064989805221558 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.7076, 0.1289, 0.6333, ..., 0.8099, 0.6262, 0.3286])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 10.064989805221558 seconds
[20.48, 20.44, 20.44, 20.48, 20.24, 20.4, 20.56, 20.44, 20.68, 20.8]
[20.72, 20.76, 24.0, 25.88, 28.52, 28.52, 29.36, 30.0, 25.88, 25.72, 23.92, 24.0, 24.2, 24.4]
14.228666067123413
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4641, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_050', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 90392, 'MATRIX_DENSITY': 9.180192695100532e-05, 'TIME_S': 10.064989805221558, 'TIME_S_1KI': 2.16871144262477, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 338.8994520378113, 'W': 23.81807615970891}
[20.48, 20.44, 20.44, 20.48, 20.24, 20.4, 20.56, 20.44, 20.68, 20.8, 20.8, 20.68, 20.84, 20.4, 20.32, 20.32, 20.36, 20.4, 20.48, 20.6]
368.81999999999994
18.440999999999995
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4641, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_050', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 90392, 'MATRIX_DENSITY': 9.180192695100532e-05, 'TIME_S': 10.064989805221558, 'TIME_S_1KI': 2.16871144262477, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 338.8994520378113, 'W': 23.81807615970891, 'J_1KI': 73.02293730614335, 'W_1KI': 5.132100012865527, 'W_D': 5.377076159708913, 'J_D': 76.5086210939885, 'W_D_1KI': 1.1586029217213776, 'J_D_1KI': 0.24964510271953838}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_055.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4732, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 10.468129873275757, "TIME_S_1KI": 2.212199888688875, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 340.41829671859733, "W": 23.891063458309905, "J_1KI": 71.93962314425134, "W_1KI": 5.048829978510124, "W_D": 5.390063458309907, "J_D": 76.80178092050545, "W_D_1KI": 1.1390666649006567, "J_D_1KI": 0.24071569418864255}

86
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_055.output Normal file
View File

@ -0,0 +1,86 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_055.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 0.2837095260620117}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.1141, 0.9590, 0.8822, ..., 0.4586, 0.3032, 0.3922])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 0.2837095260620117 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3700 -m matrices/as-caida_pruned/as-caida_G_055.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 8.209553480148315}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([1.3531e-01, 9.2471e-01, 3.7424e-01, ..., 8.7339e-04, 4.3447e-01,
4.7205e-01])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 8.209553480148315 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4732 -m matrices/as-caida_pruned/as-caida_G_055.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 10.468129873275757}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.1672, 0.2165, 0.1528, ..., 0.1782, 0.4621, 0.9393])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 10.468129873275757 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.1672, 0.2165, 0.1528, ..., 0.1782, 0.4621, 0.9393])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 10.468129873275757 seconds
[20.4, 20.44, 20.52, 20.76, 20.6, 20.84, 20.92, 20.84, 20.6, 20.6]
[20.6, 20.72, 20.72, 24.32, 26.2, 29.28, 30.44, 28.2, 28.0, 25.56, 25.68, 25.84, 25.84, 25.8]
14.24877119064331
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4732, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_055', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 91476, 'MATRIX_DENSITY': 9.290283509348351e-05, 'TIME_S': 10.468129873275757, 'TIME_S_1KI': 2.212199888688875, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 340.41829671859733, 'W': 23.891063458309905}
[20.4, 20.44, 20.52, 20.76, 20.6, 20.84, 20.92, 20.84, 20.6, 20.6, 20.64, 20.36, 20.52, 20.52, 20.48, 20.36, 20.36, 20.32, 20.44, 20.64]
370.02
18.500999999999998
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4732, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_055', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 91476, 'MATRIX_DENSITY': 9.290283509348351e-05, 'TIME_S': 10.468129873275757, 'TIME_S_1KI': 2.212199888688875, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 340.41829671859733, 'W': 23.891063458309905, 'J_1KI': 71.93962314425134, 'W_1KI': 5.048829978510124, 'W_D': 5.390063458309907, 'J_D': 76.80178092050545, 'W_D_1KI': 1.1390666649006567, 'J_D_1KI': 0.24071569418864255}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_060.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4630, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 10.542139053344727, "TIME_S_1KI": 2.2769198819319065, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 319.9495843696594, "W": 22.493385217832277, "J_1KI": 69.10358193729145, "W_1KI": 4.858182552447577, "W_D": 4.2303852178322785, "J_D": 60.173690134286886, "W_D_1KI": 0.9136901118428247, "J_D_1KI": 0.19734127685590167}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_060.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_060.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 0.2887892723083496}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.1431, 0.0707, 0.8757, ..., 0.5198, 0.2078, 0.1401])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 0.2887892723083496 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3635 -m matrices/as-caida_pruned/as-caida_G_060.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 8.24190092086792}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.3060, 0.9734, 0.9302, ..., 0.1120, 0.6990, 0.6634])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 8.24190092086792 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4630 -m matrices/as-caida_pruned/as-caida_G_060.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 10.542139053344727}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.9904, 0.2821, 0.0606, ..., 0.1141, 0.3773, 0.3266])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 10.542139053344727 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.9904, 0.2821, 0.0606, ..., 0.1141, 0.3773, 0.3266])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 10.542139053344727 seconds
[20.32, 20.48, 20.36, 20.28, 20.28, 20.0, 19.96, 19.96, 19.88, 19.96]
[20.16, 20.52, 20.88, 22.52, 24.0, 26.12, 27.04, 27.0, 26.32, 24.6, 24.6, 24.36, 24.24, 24.4]
14.224163293838501
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4630, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_060', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 94180, 'MATRIX_DENSITY': 9.564901186217454e-05, 'TIME_S': 10.542139053344727, 'TIME_S_1KI': 2.2769198819319065, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.9495843696594, 'W': 22.493385217832277}
[20.32, 20.48, 20.36, 20.28, 20.28, 20.0, 19.96, 19.96, 19.88, 19.96, 20.64, 20.48, 20.6, 20.4, 20.44, 20.52, 20.48, 20.2, 20.32, 20.32]
365.26
18.262999999999998
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4630, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_060', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 94180, 'MATRIX_DENSITY': 9.564901186217454e-05, 'TIME_S': 10.542139053344727, 'TIME_S_1KI': 2.2769198819319065, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.9495843696594, 'W': 22.493385217832277, 'J_1KI': 69.10358193729145, 'W_1KI': 4.858182552447577, 'W_D': 4.2303852178322785, 'J_D': 60.173690134286886, 'W_D_1KI': 0.9136901118428247, 'J_D_1KI': 0.19734127685590167}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_065.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4558, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 10.40420150756836, "TIME_S_1KI": 2.2826242886284245, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 329.84002380371095, "W": 23.226044906428264, "J_1KI": 72.3650776225781, "W_1KI": 5.095665841691151, "W_D": 4.6110449064282655, "J_D": 65.48283049583436, "W_D_1KI": 1.0116377591988295, "J_D_1KI": 0.2219477312853948}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_065.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_065.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 0.2867097854614258}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.2765, 0.7404, 0.5834, ..., 0.0305, 0.2184, 0.6277])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 0.2867097854614258 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3662 -m matrices/as-caida_pruned/as-caida_G_065.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 8.434634447097778}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.9579, 0.2280, 0.6543, ..., 0.1974, 0.2729, 0.8108])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 8.434634447097778 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4558 -m matrices/as-caida_pruned/as-caida_G_065.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 10.40420150756836}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.9497, 0.5746, 0.8058, ..., 0.6531, 0.4871, 0.2425])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 10.40420150756836 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.9497, 0.5746, 0.8058, ..., 0.6531, 0.4871, 0.2425])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 10.40420150756836 seconds
[20.64, 20.64, 20.6, 20.56, 20.4, 20.2, 20.16, 20.24, 20.44, 20.48]
[20.4, 20.4, 20.32, 24.04, 25.44, 28.52, 29.28, 29.6, 26.24, 25.04, 24.36, 24.48, 24.64, 24.64]
14.201299667358398
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4558, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_065', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 95068, 'MATRIX_DENSITY': 9.655086281283934e-05, 'TIME_S': 10.40420150756836, 'TIME_S_1KI': 2.2826242886284245, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 329.84002380371095, 'W': 23.226044906428264}
[20.64, 20.64, 20.6, 20.56, 20.4, 20.2, 20.16, 20.24, 20.44, 20.48, 20.76, 20.76, 20.84, 20.72, 20.92, 21.16, 21.12, 20.96, 21.04, 21.2]
372.29999999999995
18.615
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4558, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_065', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 95068, 'MATRIX_DENSITY': 9.655086281283934e-05, 'TIME_S': 10.40420150756836, 'TIME_S_1KI': 2.2826242886284245, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 329.84002380371095, 'W': 23.226044906428264, 'J_1KI': 72.3650776225781, 'W_1KI': 5.095665841691151, 'W_D': 4.6110449064282655, 'J_D': 65.48283049583436, 'W_D_1KI': 1.0116377591988295, 'J_D_1KI': 0.2219477312853948}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_070.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 5424, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 10.584465742111206, "TIME_S_1KI": 1.9514133005367267, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 328.1352438354492, "W": 23.054671486011802, "J_1KI": 60.496910736624116, "W_1KI": 4.250492530606896, "W_D": 4.752671486011803, "J_D": 67.64438252258302, "W_D_1KI": 0.8762299937337394, "J_D_1KI": 0.16154682775327053}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_070.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_070.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 0.23440051078796387}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.9209, 0.5249, 0.5745, ..., 0.6031, 0.2037, 0.8915])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 0.23440051078796387 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4479 -m matrices/as-caida_pruned/as-caida_G_070.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 8.669935703277588}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.5609, 0.8905, 0.0560, ..., 0.5802, 0.7080, 0.9443])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 8.669935703277588 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 5424 -m matrices/as-caida_pruned/as-caida_G_070.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 10.584465742111206}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.6743, 0.1388, 0.2848, ..., 0.9657, 0.9554, 0.7994])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 10.584465742111206 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.6743, 0.1388, 0.2848, ..., 0.9657, 0.9554, 0.7994])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 10.584465742111206 seconds
[20.16, 20.08, 20.12, 20.28, 20.52, 20.28, 20.4, 20.4, 20.52, 20.36]
[20.52, 20.52, 20.6, 25.24, 26.36, 28.8, 29.48, 27.2, 26.08, 23.96, 23.96, 23.84, 24.04, 24.28]
14.232917785644531
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5424, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_070', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 78684, 'MATRIX_DENSITY': 7.991130653390679e-05, 'TIME_S': 10.584465742111206, 'TIME_S_1KI': 1.9514133005367267, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 328.1352438354492, 'W': 23.054671486011802}
[20.16, 20.08, 20.12, 20.28, 20.52, 20.28, 20.4, 20.4, 20.52, 20.36, 20.28, 20.48, 20.28, 20.32, 20.36, 20.4, 20.32, 20.36, 20.32, 20.4]
366.04
18.302
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 5424, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_070', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 78684, 'MATRIX_DENSITY': 7.991130653390679e-05, 'TIME_S': 10.584465742111206, 'TIME_S_1KI': 1.9514133005367267, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 328.1352438354492, 'W': 23.054671486011802, 'J_1KI': 60.496910736624116, 'W_1KI': 4.250492530606896, 'W_D': 4.752671486011803, 'J_D': 67.64438252258302, 'W_D_1KI': 0.8762299937337394, 'J_D_1KI': 0.16154682775327053}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_075.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4325, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 10.857763290405273, "TIME_S_1KI": 2.5104655006717396, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 335.7423639297486, "W": 23.58156237869744, "J_1KI": 77.62829223809216, "W_1KI": 5.45238436501675, "W_D": 5.256562378697442, "J_D": 74.84027779102334, "W_D_1KI": 1.2153901453635703, "J_D_1KI": 0.28101506251180813}

65
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_075.output Normal file
View File

@ -0,0 +1,65 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_075.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 0.24277067184448242}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.0081, 0.1025, 0.4852, ..., 0.3080, 0.4481, 0.5761])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 0.24277067184448242 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4325 -m matrices/as-caida_pruned/as-caida_G_075.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 10.857763290405273}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.2760, 0.7522, 0.4563, ..., 0.1844, 0.1938, 0.2151])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 10.857763290405273 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.2760, 0.7522, 0.4563, ..., 0.1844, 0.1938, 0.2151])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 10.857763290405273 seconds
[20.32, 20.4, 20.16, 20.16, 20.2, 20.32, 20.32, 20.52, 20.44, 20.48]
[20.4, 20.44, 20.72, 24.96, 27.28, 28.88, 29.68, 27.12, 26.68, 26.68, 24.92, 25.16, 24.88, 24.6]
14.237494468688965
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4325, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_075', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 97492, 'MATRIX_DENSITY': 9.901267216465406e-05, 'TIME_S': 10.857763290405273, 'TIME_S_1KI': 2.5104655006717396, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 335.7423639297486, 'W': 23.58156237869744}
[20.32, 20.4, 20.16, 20.16, 20.2, 20.32, 20.32, 20.52, 20.44, 20.48, 20.2, 20.44, 20.6, 20.68, 20.64, 20.44, 20.36, 20.08, 20.16, 20.16]
366.5
18.325
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4325, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_075', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 97492, 'MATRIX_DENSITY': 9.901267216465406e-05, 'TIME_S': 10.857763290405273, 'TIME_S_1KI': 2.5104655006717396, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 335.7423639297486, 'W': 23.58156237869744, 'J_1KI': 77.62829223809216, 'W_1KI': 5.45238436501675, 'W_D': 5.256562378697442, 'J_D': 74.84027779102334, 'W_D_1KI': 1.2153901453635703, 'J_D_1KI': 0.28101506251180813}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_080.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4213, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 10.434804439544678, "TIME_S_1KI": 2.4768109279716777, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 319.78344322204595, "W": 22.486804268370133, "J_1KI": 75.90397418040493, "W_1KI": 5.337480244094501, "W_D": 3.4238042683701337, "J_D": 48.68970729637151, "W_D_1KI": 0.8126760665488093, "J_D_1KI": 0.1928972386776191}

65
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_080.output Normal file
View File

@ -0,0 +1,65 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_080.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 0.24919962882995605}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.9078, 0.2533, 0.5140, ..., 0.6043, 0.4804, 0.0746])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 0.24919962882995605 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4213 -m matrices/as-caida_pruned/as-caida_G_080.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 10.434804439544678}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.6004, 0.8484, 0.6906, ..., 0.0809, 0.2650, 0.3823])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 10.434804439544678 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.6004, 0.8484, 0.6906, ..., 0.0809, 0.2650, 0.3823])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 10.434804439544678 seconds
[20.2, 20.16, 20.88, 21.76, 22.52, 23.08, 23.36, 22.4, 21.68, 20.8]
[20.48, 20.48, 21.2, 22.52, 22.52, 25.0, 26.08, 26.52, 26.56, 25.76, 24.92, 24.96, 25.0, 25.0]
14.220937728881836
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4213, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_080', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 98112, 'MATRIX_DENSITY': 9.964234287345156e-05, 'TIME_S': 10.434804439544678, 'TIME_S_1KI': 2.4768109279716777, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.78344322204595, 'W': 22.486804268370133}
[20.2, 20.16, 20.88, 21.76, 22.52, 23.08, 23.36, 22.4, 21.68, 20.8, 20.96, 20.92, 20.32, 20.32, 20.52, 20.48, 20.4, 20.6, 20.64, 20.48]
381.26
19.063
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4213, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_080', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 98112, 'MATRIX_DENSITY': 9.964234287345156e-05, 'TIME_S': 10.434804439544678, 'TIME_S_1KI': 2.4768109279716777, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.78344322204595, 'W': 22.486804268370133, 'J_1KI': 75.90397418040493, 'W_1KI': 5.337480244094501, 'W_D': 3.4238042683701337, 'J_D': 48.68970729637151, 'W_D_1KI': 0.8126760665488093, 'J_D_1KI': 0.1928972386776191}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_085.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4299, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 10.420058727264404, "TIME_S_1KI": 2.423833153585579, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 326.41223700523375, "W": 22.905972813666082, "J_1KI": 75.92748011287131, "W_1KI": 5.32820954028055, "W_D": 4.372972813666085, "J_D": 62.31526816534997, "W_D_1KI": 1.0172069815459606, "J_D_1KI": 0.23661478984553633}

85
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_085.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_085.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 0.2965834140777588}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.4492, 0.3348, 0.2820, ..., 0.6380, 0.4778, 0.0633])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 0.2965834140777588 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3540 -m matrices/as-caida_pruned/as-caida_G_085.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 8.64511513710022}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.6270, 0.7806, 0.1009, ..., 0.0616, 0.3576, 0.1481])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 8.64511513710022 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4299 -m matrices/as-caida_pruned/as-caida_G_085.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 10.420058727264404}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.3742, 0.9296, 0.9529, ..., 0.5552, 0.6324, 0.8504])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 10.420058727264404 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.3742, 0.9296, 0.9529, ..., 0.5552, 0.6324, 0.8504])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 10.420058727264404 seconds
[21.0, 20.84, 20.84, 20.52, 20.56, 20.44, 20.52, 20.52, 20.6, 20.68]
[20.72, 20.36, 20.68, 24.16, 25.6, 27.56, 28.48, 26.48, 26.4, 24.6, 24.6, 24.68, 24.36, 24.12]
14.25009274482727
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4299, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_085', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 99166, 'MATRIX_DENSITY': 0.0001007127830784073, 'TIME_S': 10.420058727264404, 'TIME_S_1KI': 2.423833153585579, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 326.41223700523375, 'W': 22.905972813666082}
[21.0, 20.84, 20.84, 20.52, 20.56, 20.44, 20.52, 20.52, 20.6, 20.68, 20.32, 20.56, 20.4, 20.44, 20.48, 20.4, 20.72, 20.68, 20.76, 20.76]
370.65999999999997
18.532999999999998
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4299, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_085', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 99166, 'MATRIX_DENSITY': 0.0001007127830784073, 'TIME_S': 10.420058727264404, 'TIME_S_1KI': 2.423833153585579, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 326.41223700523375, 'W': 22.905972813666082, 'J_1KI': 75.92748011287131, 'W_1KI': 5.32820954028055, 'W_D': 4.372972813666085, 'J_D': 62.31526816534997, 'W_D_1KI': 1.0172069815459606, 'J_D_1KI': 0.23661478984553633}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_090.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4321, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 10.904336929321289, "TIME_S_1KI": 2.5235679077346194, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 318.7823661422729, "W": 22.454372882102824, "J_1KI": 73.77513680682085, "W_1KI": 5.19656859109068, "W_D": 4.009372882102824, "J_D": 56.92064440250394, "W_D_1KI": 0.9278807873415468, "J_D_1KI": 0.214737511534725}

110
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_090.output Normal file
View File

@ -0,0 +1,110 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_090.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 0.29803013801574707}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.5656, 0.8065, 0.8914, ..., 0.2667, 0.3733, 0.9779])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 0.29803013801574707 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3523 -m matrices/as-caida_pruned/as-caida_G_090.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 9.117663621902466}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.9346, 0.5158, 0.7987, ..., 0.7589, 0.1868, 0.6814])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 9.117663621902466 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4057 -m matrices/as-caida_pruned/as-caida_G_090.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 9.856732368469238}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.5477, 0.2740, 0.4778, ..., 0.8531, 0.3991, 0.2979])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 9.856732368469238 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4321 -m matrices/as-caida_pruned/as-caida_G_090.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 10.904336929321289}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.0196, 0.5237, 0.2833, ..., 0.4769, 0.9606, 0.8820])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 10.904336929321289 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.0196, 0.5237, 0.2833, ..., 0.4769, 0.9606, 0.8820])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 10.904336929321289 seconds
[21.0, 20.92, 21.0, 20.84, 20.44, 20.36, 20.24, 20.12, 20.04, 20.16]
[20.16, 20.08, 20.28, 21.96, 24.0, 25.96, 26.96, 26.84, 25.8, 24.72, 24.88, 24.8, 25.0, 25.0]
14.196894645690918
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4321, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_090', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 100924, 'MATRIX_DENSITY': 0.00010249820421722343, 'TIME_S': 10.904336929321289, 'TIME_S_1KI': 2.5235679077346194, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 318.7823661422729, 'W': 22.454372882102824}
[21.0, 20.92, 21.0, 20.84, 20.44, 20.36, 20.24, 20.12, 20.04, 20.16, 20.4, 20.32, 20.4, 20.48, 20.24, 20.24, 20.32, 20.56, 21.08, 21.04]
368.9
18.445
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4321, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_090', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 100924, 'MATRIX_DENSITY': 0.00010249820421722343, 'TIME_S': 10.904336929321289, 'TIME_S_1KI': 2.5235679077346194, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 318.7823661422729, 'W': 22.454372882102824, 'J_1KI': 73.77513680682085, 'W_1KI': 5.19656859109068, 'W_D': 4.009372882102824, 'J_D': 56.92064440250394, 'W_D_1KI': 0.9278807873415468, 'J_D_1KI': 0.214737511534725}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_095.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4276, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 10.67867112159729, "TIME_S_1KI": 2.4973505897093755, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 339.92240665435787, "W": 23.947333156007627, "J_1KI": 79.49541783310521, "W_1KI": 5.600405321797855, "W_D": 5.373333156007625, "J_D": 76.27222314262384, "W_D_1KI": 1.2566260888698844, "J_D_1KI": 0.2938788795299075}

89
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_095.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_095.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 0.31911325454711914}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.3207, 0.9700, 0.3262, ..., 0.9362, 0.9941, 0.3912])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 0.31911325454711914 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3290 -m matrices/as-caida_pruned/as-caida_G_095.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 8.077399730682373}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.5291, 0.8021, 0.1066, ..., 0.7881, 0.0805, 0.4870])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 8.077399730682373 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4276 -m matrices/as-caida_pruned/as-caida_G_095.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 10.67867112159729}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.1501, 0.9759, 0.0443, ..., 0.6183, 0.1649, 0.2013])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 10.67867112159729 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.1501, 0.9759, 0.0443, ..., 0.6183, 0.1649, 0.2013])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 10.67867112159729 seconds
[20.72, 20.84, 20.68, 20.36, 20.24, 20.2, 20.4, 20.32, 20.32, 20.28]
[20.32, 20.36, 20.52, 24.24, 26.84, 29.28, 30.8, 30.56, 26.92, 26.36, 25.44, 25.44, 25.36, 25.4]
14.19458293914795
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4276, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_095', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102290, 'MATRIX_DENSITY': 0.00010388551097241275, 'TIME_S': 10.67867112159729, 'TIME_S_1KI': 2.4973505897093755, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 339.92240665435787, 'W': 23.947333156007627}
[20.72, 20.84, 20.68, 20.36, 20.24, 20.2, 20.4, 20.32, 20.32, 20.28, 20.76, 20.88, 20.68, 20.84, 20.88, 21.0, 20.76, 20.88, 20.84, 20.96]
371.48
18.574
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4276, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_095', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102290, 'MATRIX_DENSITY': 0.00010388551097241275, 'TIME_S': 10.67867112159729, 'TIME_S_1KI': 2.4973505897093755, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 339.92240665435787, 'W': 23.947333156007627, 'J_1KI': 79.49541783310521, 'W_1KI': 5.600405321797855, 'W_D': 5.373333156007625, 'J_D': 76.27222314262384, 'W_D_1KI': 1.2566260888698844, 'J_D_1KI': 0.2938788795299075}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_100.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4147, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 10.192631244659424, "TIME_S_1KI": 2.4578324679670662, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 319.7802805519104, "W": 22.470932033330318, "J_1KI": 77.11123234914646, "W_1KI": 5.418599477533233, "W_D": 3.9009320333303172, "J_D": 55.51354693174359, "W_D_1KI": 0.940663620287031, "J_D_1KI": 0.226829906025327}

89
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_100.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_100.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 0.2886645793914795}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.9338, 0.7408, 0.0307, ..., 0.7058, 0.8816, 0.1464])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 0.2886645793914795 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3637 -m matrices/as-caida_pruned/as-caida_G_100.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 9.20716404914856}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.9437, 0.5183, 0.7873, ..., 0.9915, 0.6438, 0.1403])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 9.20716404914856 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4147 -m matrices/as-caida_pruned/as-caida_G_100.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 10.192631244659424}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.1025, 0.8382, 0.5877, ..., 0.7441, 0.8416, 0.8243])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 10.192631244659424 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.1025, 0.8382, 0.5877, ..., 0.7441, 0.8416, 0.8243])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 10.192631244659424 seconds
[20.28, 20.52, 20.48, 20.44, 21.0, 21.0, 21.12, 21.12, 21.12, 20.92]
[20.56, 20.52, 20.72, 21.6, 23.2, 24.8, 26.04, 26.72, 26.72, 25.36, 25.32, 25.32, 24.88, 25.04]
14.230841875076294
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4147, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_100', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102888, 'MATRIX_DENSITY': 0.00010449283852702711, 'TIME_S': 10.192631244659424, 'TIME_S_1KI': 2.4578324679670662, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.7802805519104, 'W': 22.470932033330318}
[20.28, 20.52, 20.48, 20.44, 21.0, 21.0, 21.12, 21.12, 21.12, 20.92, 20.28, 20.48, 20.24, 20.36, 20.4, 20.48, 20.6, 20.64, 20.4, 20.52]
371.40000000000003
18.57
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4147, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_100', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102888, 'MATRIX_DENSITY': 0.00010449283852702711, 'TIME_S': 10.192631244659424, 'TIME_S_1KI': 2.4578324679670662, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 319.7802805519104, 'W': 22.470932033330318, 'J_1KI': 77.11123234914646, 'W_1KI': 5.418599477533233, 'W_D': 3.9009320333303172, 'J_D': 55.51354693174359, 'W_D_1KI': 0.940663620287031, 'J_D_1KI': 0.226829906025327}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_105.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4151, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 11.793859243392944, "TIME_S_1KI": 2.8412091648742335, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 356.85992715835573, "W": 25.058368416517443, "J_1KI": 85.96962832049041, "W_1KI": 6.036706436164163, "W_D": 5.961368416517441, "J_D": 84.89672845101359, "W_D_1KI": 1.4361282622301712, "J_D_1KI": 0.3459716362876828}

89
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_105.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_105.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 0.33879852294921875}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.1080, 0.7126, 0.2741, ..., 0.0141, 0.6709, 0.0416])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 0.33879852294921875 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3099 -m matrices/as-caida_pruned/as-caida_G_105.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 7.838690519332886}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.5517, 0.4219, 0.5041, ..., 0.2191, 0.6881, 0.0206])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 7.838690519332886 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4151 -m matrices/as-caida_pruned/as-caida_G_105.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 11.793859243392944}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.6544, 0.9539, 0.4979, ..., 0.4670, 0.0344, 0.9767])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 11.793859243392944 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.6544, 0.9539, 0.4979, ..., 0.4670, 0.0344, 0.9767])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 11.793859243392944 seconds
[23.72, 23.12, 22.6, 21.2, 21.04, 21.04, 21.2, 21.6, 21.96, 22.48]
[22.96, 22.84, 22.8, 26.76, 28.44, 29.48, 30.2, 28.12, 28.12, 28.04, 26.36, 27.0, 27.12, 27.12]
14.241147756576538
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4151, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_105', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104726, 'MATRIX_DENSITY': 0.00010635950749923647, 'TIME_S': 11.793859243392944, 'TIME_S_1KI': 2.8412091648742335, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 356.85992715835573, 'W': 25.058368416517443}
[23.72, 23.12, 22.6, 21.2, 21.04, 21.04, 21.2, 21.6, 21.96, 22.48, 21.08, 20.44, 20.4, 20.6, 20.6, 20.6, 20.64, 20.48, 20.48, 20.6]
381.94
19.097
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4151, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_105', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104726, 'MATRIX_DENSITY': 0.00010635950749923647, 'TIME_S': 11.793859243392944, 'TIME_S_1KI': 2.8412091648742335, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 356.85992715835573, 'W': 25.058368416517443, 'J_1KI': 85.96962832049041, 'W_1KI': 6.036706436164163, 'W_D': 5.961368416517441, 'J_D': 84.89672845101359, 'W_D_1KI': 1.4361282622301712, 'J_D_1KI': 0.3459716362876828}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_110.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4044, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 10.143035173416138, "TIME_S_1KI": 2.5081689350682836, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 330.16996871948237, "W": 23.153445351321693, "J_1KI": 81.6444037387444, "W_1KI": 5.725382134352545, "W_D": 4.751445351321696, "J_D": 67.7559878978729, "W_D_1KI": 1.174937030494979, "J_D_1KI": 0.29053833592853096}

68
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_110.output Normal file
View File

@ -0,0 +1,68 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_110.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 0.25959014892578125}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.1398, 0.1495, 0.6361, ..., 0.4484, 0.6307, 0.8400])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 0.25959014892578125 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4044 -m matrices/as-caida_pruned/as-caida_G_110.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 10.143035173416138}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.3022, 0.4527, 0.1551, ..., 0.4286, 0.4985, 0.4790])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 10.143035173416138 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.3022, 0.4527, 0.1551, ..., 0.4286, 0.4985, 0.4790])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 10.143035173416138 seconds
[20.64, 20.72, 20.32, 20.56, 20.36, 20.32, 20.52, 21.16, 21.16, 21.04]
[21.28, 21.2, 20.76, 24.0, 26.08, 28.36, 29.2, 26.96, 26.92, 24.32, 24.24, 24.24, 24.4, 24.8]
14.260079383850098
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4044, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_110', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104846, 'MATRIX_DENSITY': 0.0001064813792493263, 'TIME_S': 10.143035173416138, 'TIME_S_1KI': 2.5081689350682836, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 330.16996871948237, 'W': 23.153445351321693}
[20.64, 20.72, 20.32, 20.56, 20.36, 20.32, 20.52, 21.16, 21.16, 21.04, 20.6, 20.2, 20.24, 19.92, 20.12, 20.24, 20.08, 20.32, 20.44, 20.44]
368.03999999999996
18.401999999999997
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4044, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_110', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104846, 'MATRIX_DENSITY': 0.0001064813792493263, 'TIME_S': 10.143035173416138, 'TIME_S_1KI': 2.5081689350682836, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 330.16996871948237, 'W': 23.153445351321693, 'J_1KI': 81.6444037387444, 'W_1KI': 5.725382134352545, 'W_D': 4.751445351321696, 'J_D': 67.7559878978729, 'W_D_1KI': 1.174937030494979, 'J_D_1KI': 0.29053833592853096}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_115.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 3898, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 10.276582956314087, "TIME_S_1KI": 2.6363732571354768, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 299.2206564903259, "W": 22.616425322779282, "J_1KI": 76.76261069531193, "W_1KI": 5.802058830882319, "W_D": 4.17842532277928, "J_D": 55.28155534458152, "W_D_1KI": 1.0719408216468136, "J_D_1KI": 0.27499764536860277}

89
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_115.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_115.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 0.29033493995666504}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.0287, 0.9302, 0.4533, ..., 0.1887, 0.8093, 0.0476])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 0.29033493995666504 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3616 -m matrices/as-caida_pruned/as-caida_G_115.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 9.739661455154419}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.1589, 0.7243, 0.3638, ..., 0.5413, 0.9750, 0.3668])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 9.739661455154419 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3898 -m matrices/as-caida_pruned/as-caida_G_115.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 10.276582956314087}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.5554, 0.1424, 0.7572, ..., 0.7612, 0.5304, 0.9292])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 10.276582956314087 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.5554, 0.1424, 0.7572, ..., 0.7612, 0.5304, 0.9292])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 10.276582956314087 seconds
[20.68, 20.84, 20.88, 20.88, 20.48, 20.32, 20.28, 20.04, 20.2, 20.24]
[20.12, 20.44, 21.0, 22.68, 25.16, 26.56, 26.56, 27.28, 26.48, 26.6, 24.04, 24.28, 24.76]
13.230236530303955
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 3898, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_115', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106312, 'MATRIX_DENSITY': 0.00010797024579625715, 'TIME_S': 10.276582956314087, 'TIME_S_1KI': 2.6363732571354768, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 299.2206564903259, 'W': 22.616425322779282}
[20.68, 20.84, 20.88, 20.88, 20.48, 20.32, 20.28, 20.04, 20.2, 20.24, 20.4, 20.36, 20.36, 20.48, 20.52, 20.4, 20.52, 20.52, 20.68, 20.68]
368.76000000000005
18.438000000000002
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 3898, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_115', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106312, 'MATRIX_DENSITY': 0.00010797024579625715, 'TIME_S': 10.276582956314087, 'TIME_S_1KI': 2.6363732571354768, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 299.2206564903259, 'W': 22.616425322779282, 'J_1KI': 76.76261069531193, 'W_1KI': 5.802058830882319, 'W_D': 4.17842532277928, 'J_D': 55.28155534458152, 'W_D_1KI': 1.0719408216468136, 'J_D_1KI': 0.27499764536860277}

1
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_120.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 4027, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 10.212517261505127, "TIME_S_1KI": 2.5360112395095924, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 332.28165603637694, "W": 23.272101675798375, "J_1KI": 82.51344823351799, "W_1KI": 5.779017053836199, "W_D": 5.076101675798377, "J_D": 72.47714428806304, "W_D_1KI": 1.2605169296742929, "J_D_1KI": 0.3130163719081929}

89
pytorch/output_as-caida_16core/altra_16_csr_10_10_10_as-caida_G_120.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 100 -m matrices/as-caida_pruned/as-caida_G_120.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 0.311673641204834}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.7695, 0.0539, 0.1056, ..., 0.2425, 0.4084, 0.5084])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 0.311673641204834 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 3368 -m matrices/as-caida_pruned/as-caida_G_120.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 8.780853271484375}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.1115, 0.2867, 0.9875, ..., 0.7692, 0.4068, 0.8761])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 8.780853271484375 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 4027 -m matrices/as-caida_pruned/as-caida_G_120.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 10.212517261505127}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.2551, 0.3437, 0.0774, ..., 0.8176, 0.1852, 0.3451])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 10.212517261505127 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at /space/jenkins/workspace/Releases/pytorch-dls/pytorch-dls/aten/src/ATen/SparseCsrTensorImpl.cpp:55.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.2551, 0.3437, 0.0774, ..., 0.8176, 0.1852, 0.3451])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 10.212517261505127 seconds
[20.0, 20.04, 19.96, 20.28, 20.6, 20.64, 20.72, 20.56, 20.44, 20.44]
[20.44, 20.44, 20.56, 24.2, 25.36, 27.88, 29.04, 29.4, 26.08, 25.44, 24.64, 24.64, 24.96, 24.96]
14.278111219406128
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4027, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_120', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106510, 'MATRIX_DENSITY': 0.0001081713341839054, 'TIME_S': 10.212517261505127, 'TIME_S_1KI': 2.5360112395095924, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 332.28165603637694, 'W': 23.272101675798375}
[20.0, 20.04, 19.96, 20.28, 20.6, 20.64, 20.72, 20.56, 20.44, 20.44, 19.8, 19.8, 19.92, 19.96, 20.04, 20.24, 20.24, 20.08, 20.12, 20.32]
363.91999999999996
18.195999999999998
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 4027, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_120', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106510, 'MATRIX_DENSITY': 0.0001081713341839054, 'TIME_S': 10.212517261505127, 'TIME_S_1KI': 2.5360112395095924, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 332.28165603637694, 'W': 23.272101675798375, 'J_1KI': 82.51344823351799, 'W_1KI': 5.779017053836199, 'W_D': 5.076101675798377, 'J_D': 72.47714428806304, 'W_D_1KI': 1.2605169296742929, 'J_D_1KI': 0.3130163719081929}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_005.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 130629, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 11.046883583068848, "TIME_S_1KI": 0.08456685409111948, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1447.5023779034616, "W": 103.45, "J_1KI": 11.081018593906878, "W_1KI": 0.7919374717711994, "W_D": 67.28425, "J_D": 941.4607237356305, "W_D_1KI": 0.5150789640891379, "J_D_1KI": 0.003943067497180089}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_005.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 0.02265310287475586}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.1109, 0.3688, 0.8394, ..., 0.5778, 0.7474, 0.0459])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 0.02265310287475586 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '46351', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 3.7256975173950195}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.7624, 0.1488, 0.7288, ..., 0.4517, 0.7426, 0.4871])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 3.7256975173950195 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '130629', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 11.046883583068848}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.6772, 0.5623, 0.5296, ..., 0.8301, 0.9620, 0.4995])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 11.046883583068848 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.6772, 0.5623, 0.5296, ..., 0.8301, 0.9620, 0.4995])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 11.046883583068848 seconds
[40.01, 39.79, 39.68, 39.98, 39.67, 39.17, 44.96, 39.73, 39.43, 39.12]
[103.45]
13.992289781570435
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 130629, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 11.046883583068848, 'TIME_S_1KI': 0.08456685409111948, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1447.5023779034616, 'W': 103.45}
[40.01, 39.79, 39.68, 39.98, 39.67, 39.17, 44.96, 39.73, 39.43, 39.12, 40.63, 39.49, 44.41, 39.16, 39.85, 39.65, 39.59, 39.3, 39.78, 39.59]
723.315
36.16575
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 130629, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 11.046883583068848, 'TIME_S_1KI': 0.08456685409111948, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1447.5023779034616, 'W': 103.45, 'J_1KI': 11.081018593906878, 'W_1KI': 0.7919374717711994, 'W_D': 67.28425, 'J_D': 941.4607237356305, 'W_D_1KI': 0.5150789640891379, 'J_D_1KI': 0.003943067497180089}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_010.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120346, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 10.589650392532349, "TIME_S_1KI": 0.08799337238073844, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1358.952927970886, "W": 103.07, "J_1KI": 11.292048991830937, "W_1KI": 0.856447243780433, "W_D": 67.06599999999999, "J_D": 884.2489285659789, "W_D_1KI": 0.5572765193691521, "J_D_1KI": 0.004630619375543451}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_010.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_010.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 0.028674840927124023}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.9587, 0.4811, 0.8625, ..., 0.9129, 0.1225, 0.6838])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 0.028674840927124023 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36617', '-m', 'matrices/as-caida_pruned/as-caida_G_010.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 3.194762945175171}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.6315, 0.1026, 0.7910, ..., 0.8210, 0.9570, 0.6355])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 3.194762945175171 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120346', '-m', 'matrices/as-caida_pruned/as-caida_G_010.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 10.589650392532349}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.5481, 0.8859, 0.1543, ..., 0.7989, 0.5471, 0.0598])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 10.589650392532349 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 28, 28, ..., 74993, 74993, 74994]),
col_indices=tensor([ 1040, 2020, 2054, ..., 160, 160, 12170]),
values=tensor([1., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=74994, layout=torch.sparse_csr)
tensor([0.5481, 0.8859, 0.1543, ..., 0.7989, 0.5471, 0.0598])
Matrix Type: SuiteSparse
Matrix: as-caida_G_010
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 74994
Density: 7.616375021864427e-05
Time: 10.589650392532349 seconds
[40.32, 39.72, 39.52, 39.19, 39.62, 39.54, 39.3, 39.53, 39.19, 44.46]
[103.07]
13.184757232666016
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120346, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_010', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 74994, 'MATRIX_DENSITY': 7.616375021864427e-05, 'TIME_S': 10.589650392532349, 'TIME_S_1KI': 0.08799337238073844, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1358.952927970886, 'W': 103.07}
[40.32, 39.72, 39.52, 39.19, 39.62, 39.54, 39.3, 39.53, 39.19, 44.46, 39.95, 39.65, 39.59, 39.68, 39.55, 39.32, 45.78, 39.25, 39.47, 39.63]
720.0800000000002
36.004000000000005
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120346, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_010', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 74994, 'MATRIX_DENSITY': 7.616375021864427e-05, 'TIME_S': 10.589650392532349, 'TIME_S_1KI': 0.08799337238073844, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1358.952927970886, 'W': 103.07, 'J_1KI': 11.292048991830937, 'W_1KI': 0.856447243780433, 'W_D': 67.06599999999999, 'J_D': 884.2489285659789, 'W_D_1KI': 0.5572765193691521, 'J_D_1KI': 0.004630619375543451}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_015.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 122232, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 10.012900829315186, "TIME_S_1KI": 0.08191718068357864, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1344.7570972156525, "W": 103.87, "J_1KI": 11.001677933893355, "W_1KI": 0.8497774723476668, "W_D": 67.6435, "J_D": 875.7492702946663, "W_D_1KI": 0.5534025459781399, "J_D_1KI": 0.004527476814403265}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_015.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_015.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 0.02150440216064453}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.2547, 0.3736, 0.4755, ..., 0.0068, 0.6237, 0.5320])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 0.02150440216064453 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '48827', '-m', 'matrices/as-caida_pruned/as-caida_G_015.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 4.194327354431152}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.7417, 0.4748, 0.0091, ..., 0.5058, 0.6479, 0.7190])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 4.194327354431152 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '122232', '-m', 'matrices/as-caida_pruned/as-caida_G_015.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_015", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 77124, "MATRIX_DENSITY": 7.832697378273889e-05, "TIME_S": 10.012900829315186}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.8581, 0.1137, 0.8207, ..., 0.0910, 0.4048, 0.6394])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 10.012900829315186 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 4, 4, ..., 77124, 77124, 77124]),
col_indices=tensor([1040, 2054, 4842, ..., 160, 160, 8230]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=77124, layout=torch.sparse_csr)
tensor([0.8581, 0.1137, 0.8207, ..., 0.0910, 0.4048, 0.6394])
Matrix Type: SuiteSparse
Matrix: as-caida_G_015
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 77124
Density: 7.832697378273889e-05
Time: 10.012900829315186 seconds
[40.42, 39.32, 39.31, 39.25, 54.33, 39.25, 39.4, 39.12, 39.27, 40.74]
[103.87]
12.946539878845215
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122232, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_015', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 77124, 'MATRIX_DENSITY': 7.832697378273889e-05, 'TIME_S': 10.012900829315186, 'TIME_S_1KI': 0.08191718068357864, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1344.7570972156525, 'W': 103.87}
[40.42, 39.32, 39.31, 39.25, 54.33, 39.25, 39.4, 39.12, 39.27, 40.74, 39.91, 39.57, 39.25, 39.28, 39.45, 39.43, 39.72, 39.1, 39.4, 39.09]
724.53
36.2265
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122232, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_015', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 77124, 'MATRIX_DENSITY': 7.832697378273889e-05, 'TIME_S': 10.012900829315186, 'TIME_S_1KI': 0.08191718068357864, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1344.7570972156525, 'W': 103.87, 'J_1KI': 11.001677933893355, 'W_1KI': 0.8497774723476668, 'W_D': 67.6435, 'J_D': 875.7492702946663, 'W_D_1KI': 0.5534025459781399, 'J_D_1KI': 0.004527476814403265}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_020.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 117384, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 10.323657274246216, "TIME_S_1KI": 0.08794773797320092, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1313.2328101539613, "W": 103.76, "J_1KI": 11.187494123168074, "W_1KI": 0.8839364819736932, "W_D": 68.35725000000002, "J_D": 865.1598256736401, "W_D_1KI": 0.5823387344101413, "J_D_1KI": 0.0049609719758241435}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_020.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_020.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 0.027828693389892578}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.1453, 0.8777, 0.6885, ..., 0.8576, 0.3773, 0.4559])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 0.027828693389892578 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '37730', '-m', 'matrices/as-caida_pruned/as-caida_G_020.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 3.3749337196350098}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.1719, 0.8159, 0.9898, ..., 0.9766, 0.3485, 0.4418])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 3.3749337196350098 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '117384', '-m', 'matrices/as-caida_pruned/as-caida_G_020.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_020", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 80948, "MATRIX_DENSITY": 8.221062021893506e-05, "TIME_S": 10.323657274246216}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.6104, 0.1780, 0.2799, ..., 0.5188, 0.3967, 0.6247])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 10.323657274246216 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 80944, 80946, 80948]),
col_indices=tensor([ 1040, 5699, 106, ..., 31378, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=80948, layout=torch.sparse_csr)
tensor([0.6104, 0.1780, 0.2799, ..., 0.5188, 0.3967, 0.6247])
Matrix Type: SuiteSparse
Matrix: as-caida_G_020
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 80948
Density: 8.221062021893506e-05
Time: 10.323657274246216 seconds
[41.71, 39.1, 39.14, 39.11, 39.01, 39.07, 39.54, 39.01, 39.19, 39.4]
[103.76]
12.656445741653442
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117384, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_020', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 80948, 'MATRIX_DENSITY': 8.221062021893506e-05, 'TIME_S': 10.323657274246216, 'TIME_S_1KI': 0.08794773797320092, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1313.2328101539613, 'W': 103.76}
[41.71, 39.1, 39.14, 39.11, 39.01, 39.07, 39.54, 39.01, 39.19, 39.4, 40.11, 39.48, 39.53, 39.22, 39.48, 39.26, 39.66, 39.01, 39.15, 38.97]
708.0549999999998
35.40274999999999
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117384, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_020', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 80948, 'MATRIX_DENSITY': 8.221062021893506e-05, 'TIME_S': 10.323657274246216, 'TIME_S_1KI': 0.08794773797320092, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1313.2328101539613, 'W': 103.76, 'J_1KI': 11.187494123168074, 'W_1KI': 0.8839364819736932, 'W_D': 68.35725000000002, 'J_D': 865.1598256736401, 'W_D_1KI': 0.5823387344101413, 'J_D_1KI': 0.0049609719758241435}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_025.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120118, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 12.039484739303589, "TIME_S_1KI": 0.10023047952266595, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1377.8382610416413, "W": 104.04, "J_1KI": 11.47070598113223, "W_1KI": 0.8661482875172747, "W_D": 68.01325, "J_D": 900.7233574374318, "W_D_1KI": 0.5662203000382956, "J_D_1KI": 0.004713867197574848}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_025.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_025.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 0.028760671615600586}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.0546, 0.6478, 0.5019, ..., 0.1774, 0.5884, 0.7696])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 0.028760671615600586 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36508', '-m', 'matrices/as-caida_pruned/as-caida_G_025.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 3.191291332244873}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.7513, 0.6718, 0.2286, ..., 0.8031, 0.6348, 0.1488])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 3.191291332244873 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120118', '-m', 'matrices/as-caida_pruned/as-caida_G_025.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_025", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 85850, "MATRIX_DENSITY": 8.718908121010495e-05, "TIME_S": 12.039484739303589}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.0727, 0.2026, 0.8265, ..., 0.2293, 0.8547, 0.4127])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 12.039484739303589 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 3, 3, ..., 85845, 85847, 85850]),
col_indices=tensor([ 346, 13811, 21783, ..., 15310, 17998, 31377]),
values=tensor([1., 1., 1., ..., 1., 1., 3.]), size=(31379, 31379),
nnz=85850, layout=torch.sparse_csr)
tensor([0.0727, 0.2026, 0.8265, ..., 0.2293, 0.8547, 0.4127])
Matrix Type: SuiteSparse
Matrix: as-caida_G_025
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 85850
Density: 8.718908121010495e-05
Time: 12.039484739303589 seconds
[41.14, 39.71, 39.25, 39.89, 39.22, 39.19, 39.32, 39.24, 44.24, 39.23]
[104.04]
13.243351221084595
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120118, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_025', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 85850, 'MATRIX_DENSITY': 8.718908121010495e-05, 'TIME_S': 12.039484739303589, 'TIME_S_1KI': 0.10023047952266595, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1377.8382610416413, 'W': 104.04}
[41.14, 39.71, 39.25, 39.89, 39.22, 39.19, 39.32, 39.24, 44.24, 39.23, 40.57, 39.23, 39.41, 39.27, 39.7, 44.55, 39.47, 39.31, 39.26, 39.61]
720.5350000000001
36.02675000000001
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120118, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_025', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 85850, 'MATRIX_DENSITY': 8.718908121010495e-05, 'TIME_S': 12.039484739303589, 'TIME_S_1KI': 0.10023047952266595, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1377.8382610416413, 'W': 104.04, 'J_1KI': 11.47070598113223, 'W_1KI': 0.8661482875172747, 'W_D': 68.01325, 'J_D': 900.7233574374318, 'W_D_1KI': 0.5662203000382956, 'J_D_1KI': 0.004713867197574848}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_030.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120261, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 10.17577338218689, "TIME_S_1KI": 0.08461407590313476, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1384.515518093109, "W": 104.29999999999998, "J_1KI": 11.512589435420535, "W_1KI": 0.8672803319446868, "W_D": 68.51474999999998, "J_D": 909.4893057839868, "W_D_1KI": 0.5697171152742783, "J_D_1KI": 0.004737338915145211}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_030.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_030.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 0.028615236282348633}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.8200, 0.5639, 0.4277, ..., 0.0741, 0.1847, 0.6824])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 0.028615236282348633 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36693', '-m', 'matrices/as-caida_pruned/as-caida_G_030.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 3.2036514282226562}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.5381, 0.7793, 0.3349, ..., 0.9005, 0.4954, 0.1180])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 3.2036514282226562 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120261', '-m', 'matrices/as-caida_pruned/as-caida_G_030.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_030", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 86850, "MATRIX_DENSITY": 8.820467912752026e-05, "TIME_S": 10.17577338218689}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.6109, 0.7009, 0.2057, ..., 0.1567, 0.4641, 0.2303])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 10.17577338218689 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 86850, 86850, 86850]),
col_indices=tensor([ 1809, 21783, 106, ..., 7018, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=86850, layout=torch.sparse_csr)
tensor([0.6109, 0.7009, 0.2057, ..., 0.1567, 0.4641, 0.2303])
Matrix Type: SuiteSparse
Matrix: as-caida_G_030
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 86850
Density: 8.820467912752026e-05
Time: 10.17577338218689 seconds
[42.34, 39.73, 39.59, 39.75, 39.28, 39.48, 39.88, 39.38, 40.44, 39.3]
[104.3]
13.274357795715332
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120261, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_030', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 86850, 'MATRIX_DENSITY': 8.820467912752026e-05, 'TIME_S': 10.17577338218689, 'TIME_S_1KI': 0.08461407590313476, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1384.515518093109, 'W': 104.29999999999998}
[42.34, 39.73, 39.59, 39.75, 39.28, 39.48, 39.88, 39.38, 40.44, 39.3, 40.72, 39.38, 39.58, 39.29, 39.9, 40.41, 39.95, 39.32, 39.36, 39.61]
715.705
35.785250000000005
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120261, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_030', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 86850, 'MATRIX_DENSITY': 8.820467912752026e-05, 'TIME_S': 10.17577338218689, 'TIME_S_1KI': 0.08461407590313476, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1384.515518093109, 'W': 104.29999999999998, 'J_1KI': 11.512589435420535, 'W_1KI': 0.8672803319446868, 'W_D': 68.51474999999998, 'J_D': 909.4893057839868, 'W_D_1KI': 0.5697171152742783, 'J_D_1KI': 0.004737338915145211}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_035.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 117637, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 10.62736988067627, "TIME_S_1KI": 0.09034036808721975, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1348.652883746624, "W": 104.36999999999999, "J_1KI": 11.464529729138144, "W_1KI": 0.8872208573833061, "W_D": 68.05524999999999, "J_D": 879.3993404867051, "W_D_1KI": 0.5785190883820566, "J_D_1KI": 0.00491783272594555}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_035.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_035.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 0.029236316680908203}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.3402, 0.2499, 0.3172, ..., 0.1488, 0.4072, 0.2122])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 0.029236316680908203 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '35914', '-m', 'matrices/as-caida_pruned/as-caida_G_035.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 3.2055723667144775}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.5217, 0.6973, 0.2862, ..., 0.9929, 0.5018, 0.4794])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 3.2055723667144775 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '117637', '-m', 'matrices/as-caida_pruned/as-caida_G_035.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_035", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 87560, "MATRIX_DENSITY": 8.892575364888514e-05, "TIME_S": 10.62736988067627}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.6504, 0.2249, 0.2739, ..., 0.8117, 0.7999, 0.0068])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 10.62736988067627 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 2, 2, ..., 87559, 87559, 87560]),
col_indices=tensor([ 1809, 21783, 106, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=87560, layout=torch.sparse_csr)
tensor([0.6504, 0.2249, 0.2739, ..., 0.8117, 0.7999, 0.0068])
Matrix Type: SuiteSparse
Matrix: as-caida_G_035
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 87560
Density: 8.892575364888514e-05
Time: 10.62736988067627 seconds
[39.98, 40.1, 39.44, 39.73, 39.36, 44.62, 40.72, 39.73, 40.41, 39.38]
[104.37]
12.921844244003296
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117637, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_035', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 87560, 'MATRIX_DENSITY': 8.892575364888514e-05, 'TIME_S': 10.62736988067627, 'TIME_S_1KI': 0.09034036808721975, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1348.652883746624, 'W': 104.36999999999999}
[39.98, 40.1, 39.44, 39.73, 39.36, 44.62, 40.72, 39.73, 40.41, 39.38, 40.08, 39.64, 42.5, 41.53, 39.76, 39.42, 39.32, 39.21, 40.44, 41.29]
726.2950000000001
36.314750000000004
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117637, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_035', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 87560, 'MATRIX_DENSITY': 8.892575364888514e-05, 'TIME_S': 10.62736988067627, 'TIME_S_1KI': 0.09034036808721975, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1348.652883746624, 'W': 104.36999999999999, 'J_1KI': 11.464529729138144, 'W_1KI': 0.8872208573833061, 'W_D': 68.05524999999999, 'J_D': 879.3993404867051, 'W_D_1KI': 0.5785190883820566, 'J_D_1KI': 0.00491783272594555}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_040.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120015, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 11.120282411575317, "TIME_S_1KI": 0.09265743791672139, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1376.9933811092376, "W": 104.52, "J_1KI": 11.473510653745262, "W_1KI": 0.870891138607674, "W_D": 68.45974999999999, "J_D": 901.9194663451312, "W_D_1KI": 0.5704266133399991, "J_D_1KI": 0.004752960991042779}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_040.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_040.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 0.02805638313293457}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.1538, 0.8881, 0.1064, ..., 0.9482, 0.5337, 0.0273])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 0.02805638313293457 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '37424', '-m', 'matrices/as-caida_pruned/as-caida_G_040.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 3.2741847038269043}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.4384, 0.9373, 0.4871, ..., 0.2356, 0.2534, 0.8446])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 3.2741847038269043 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120015', '-m', 'matrices/as-caida_pruned/as-caida_G_040.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_040", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89658, "MATRIX_DENSITY": 9.105647807962247e-05, "TIME_S": 11.120282411575317}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.7128, 0.1137, 0.0056, ..., 0.4103, 0.4063, 0.8720])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 11.120282411575317 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89657, 89657, 89658]),
col_indices=tensor([ 106, 329, 1040, ..., 10144, 882, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89658, layout=torch.sparse_csr)
tensor([0.7128, 0.1137, 0.0056, ..., 0.4103, 0.4063, 0.8720])
Matrix Type: SuiteSparse
Matrix: as-caida_G_040
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89658
Density: 9.105647807962247e-05
Time: 11.120282411575317 seconds
[41.64, 39.35, 39.52, 39.43, 39.11, 39.72, 39.25, 39.84, 44.65, 39.54]
[104.52]
13.174448728561401
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120015, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_040', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89658, 'MATRIX_DENSITY': 9.105647807962247e-05, 'TIME_S': 11.120282411575317, 'TIME_S_1KI': 0.09265743791672139, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1376.9933811092376, 'W': 104.52}
[41.64, 39.35, 39.52, 39.43, 39.11, 39.72, 39.25, 39.84, 44.65, 39.54, 40.23, 39.89, 39.33, 39.08, 39.12, 44.38, 39.71, 39.24, 39.35, 39.06]
721.205
36.06025
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120015, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_040', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89658, 'MATRIX_DENSITY': 9.105647807962247e-05, 'TIME_S': 11.120282411575317, 'TIME_S_1KI': 0.09265743791672139, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1376.9933811092376, 'W': 104.52, 'J_1KI': 11.473510653745262, 'W_1KI': 0.870891138607674, 'W_D': 68.45974999999999, 'J_D': 901.9194663451312, 'W_D_1KI': 0.5704266133399991, 'J_D_1KI': 0.004752960991042779}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_045.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 129597, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 11.245419979095459, "TIME_S_1KI": 0.08677222450439022, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1476.8643749785424, "W": 104.97, "J_1KI": 11.395822241090013, "W_1KI": 0.8099724530660432, "W_D": 69.20224999999999, "J_D": 973.6337781590818, "W_D_1KI": 0.5339803390510582, "J_D_1KI": 0.004120314043157312}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_045.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_045.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 0.027070283889770508}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.9579, 0.3823, 0.7927, ..., 0.9257, 0.1735, 0.9344])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 0.027070283889770508 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '38787', '-m', 'matrices/as-caida_pruned/as-caida_G_045.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 3.142535448074341}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.7648, 0.3979, 0.1181, ..., 0.8603, 0.9960, 0.9728])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 3.142535448074341 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '129597', '-m', 'matrices/as-caida_pruned/as-caida_G_045.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_045", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 89152, "MATRIX_DENSITY": 9.054258553341032e-05, "TIME_S": 11.245419979095459}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.1180, 0.3748, 0.1643, ..., 0.9664, 0.3966, 0.4847])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 11.245419979095459 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 89150, 89150, 89152]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=89152, layout=torch.sparse_csr)
tensor([0.1180, 0.3748, 0.1643, ..., 0.9664, 0.3966, 0.4847])
Matrix Type: SuiteSparse
Matrix: as-caida_G_045
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 89152
Density: 9.054258553341032e-05
Time: 11.245419979095459 seconds
[39.97, 39.69, 41.18, 39.43, 39.33, 39.4, 39.92, 39.67, 40.11, 39.3]
[104.97]
14.069394826889038
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 129597, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_045', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89152, 'MATRIX_DENSITY': 9.054258553341032e-05, 'TIME_S': 11.245419979095459, 'TIME_S_1KI': 0.08677222450439022, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1476.8643749785424, 'W': 104.97}
[39.97, 39.69, 41.18, 39.43, 39.33, 39.4, 39.92, 39.67, 40.11, 39.3, 40.3, 39.27, 39.34, 39.54, 39.86, 39.21, 39.39, 41.12, 39.24, 39.74]
715.355
35.76775
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 129597, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_045', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 89152, 'MATRIX_DENSITY': 9.054258553341032e-05, 'TIME_S': 11.245419979095459, 'TIME_S_1KI': 0.08677222450439022, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1476.8643749785424, 'W': 104.97, 'J_1KI': 11.395822241090013, 'W_1KI': 0.8099724530660432, 'W_D': 69.20224999999999, 'J_D': 973.6337781590818, 'W_D_1KI': 0.5339803390510582, 'J_D_1KI': 0.004120314043157312}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_050.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 117251, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 10.673714637756348, "TIME_S_1KI": 0.09103303714046232, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1359.178429517746, "W": 103.34, "J_1KI": 11.59204125779521, "W_1KI": 0.8813570886389029, "W_D": 67.419, "J_D": 886.7277969775199, "W_D_1KI": 0.5749972281686296, "J_D_1KI": 0.004903985707317035}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_050.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_050.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 0.028463363647460938}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.6452, 0.5948, 0.0594, ..., 0.8774, 0.4998, 0.0874])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 0.028463363647460938 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36889', '-m', 'matrices/as-caida_pruned/as-caida_G_050.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 3.303457498550415}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.0413, 0.9306, 0.8847, ..., 0.0971, 0.6839, 0.6839])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 3.303457498550415 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '117251', '-m', 'matrices/as-caida_pruned/as-caida_G_050.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_050", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 90392, "MATRIX_DENSITY": 9.180192695100532e-05, "TIME_S": 10.673714637756348}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.1911, 0.5487, 0.3865, ..., 0.4028, 0.6858, 0.4918])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 10.673714637756348 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 90390, 90390, 90392]),
col_indices=tensor([ 5326, 106, 329, ..., 882, 2232, 16085]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=90392, layout=torch.sparse_csr)
tensor([0.1911, 0.5487, 0.3865, ..., 0.4028, 0.6858, 0.4918])
Matrix Type: SuiteSparse
Matrix: as-caida_G_050
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 90392
Density: 9.180192695100532e-05
Time: 10.673714637756348 seconds
[40.43, 39.16, 39.11, 39.49, 39.45, 39.08, 39.11, 38.88, 44.34, 39.11]
[103.34]
13.152491092681885
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117251, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_050', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 90392, 'MATRIX_DENSITY': 9.180192695100532e-05, 'TIME_S': 10.673714637756348, 'TIME_S_1KI': 0.09103303714046232, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1359.178429517746, 'W': 103.34}
[40.43, 39.16, 39.11, 39.49, 39.45, 39.08, 39.11, 38.88, 44.34, 39.11, 40.07, 39.08, 39.26, 39.54, 41.34, 43.45, 39.2, 39.4, 39.22, 39.01]
718.4200000000001
35.92100000000001
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 117251, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_050', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 90392, 'MATRIX_DENSITY': 9.180192695100532e-05, 'TIME_S': 10.673714637756348, 'TIME_S_1KI': 0.09103303714046232, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1359.178429517746, 'W': 103.34, 'J_1KI': 11.59204125779521, 'W_1KI': 0.8813570886389029, 'W_D': 67.419, 'J_D': 886.7277969775199, 'W_D_1KI': 0.5749972281686296, 'J_D_1KI': 0.004903985707317035}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_055.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 122142, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 11.075148582458496, "TIME_S_1KI": 0.09067436739580567, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1444.9410432815553, "W": 104.47000000000001, "J_1KI": 11.830009687753233, "W_1KI": 0.855315943737617, "W_D": 68.78025000000001, "J_D": 951.3104833173753, "W_D_1KI": 0.5631171095937516, "J_D_1KI": 0.004610347870460215}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_055.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_055.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 0.02579522132873535}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.4101, 0.8209, 0.7582, ..., 0.7042, 0.4089, 0.9250])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 0.02579522132873535 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '40705', '-m', 'matrices/as-caida_pruned/as-caida_G_055.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 3.499220609664917}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.5165, 0.5886, 0.2570, ..., 0.5351, 0.5985, 0.2855])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 3.499220609664917 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '122142', '-m', 'matrices/as-caida_pruned/as-caida_G_055.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_055", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 91476, "MATRIX_DENSITY": 9.290283509348351e-05, "TIME_S": 11.075148582458496}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.1241, 0.2147, 0.8696, ..., 0.1307, 0.0728, 0.1644])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 11.075148582458496 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 1, ..., 91475, 91475, 91476]),
col_indices=tensor([21783, 106, 329, ..., 160, 882, 17255]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=91476, layout=torch.sparse_csr)
tensor([0.1241, 0.2147, 0.8696, ..., 0.1307, 0.0728, 0.1644])
Matrix Type: SuiteSparse
Matrix: as-caida_G_055
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 91476
Density: 9.290283509348351e-05
Time: 11.075148582458496 seconds
[40.03, 39.33, 39.47, 39.33, 39.45, 39.87, 39.7, 39.82, 40.96, 39.66]
[104.47]
13.831157684326172
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122142, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_055', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 91476, 'MATRIX_DENSITY': 9.290283509348351e-05, 'TIME_S': 11.075148582458496, 'TIME_S_1KI': 0.09067436739580567, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1444.9410432815553, 'W': 104.47000000000001}
[40.03, 39.33, 39.47, 39.33, 39.45, 39.87, 39.7, 39.82, 40.96, 39.66, 39.88, 39.36, 39.59, 39.25, 40.14, 39.15, 39.32, 39.55, 39.92, 39.6]
713.7950000000001
35.689750000000004
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122142, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_055', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 91476, 'MATRIX_DENSITY': 9.290283509348351e-05, 'TIME_S': 11.075148582458496, 'TIME_S_1KI': 0.09067436739580567, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1444.9410432815553, 'W': 104.47000000000001, 'J_1KI': 11.830009687753233, 'W_1KI': 0.855315943737617, 'W_D': 68.78025000000001, 'J_D': 951.3104833173753, 'W_D_1KI': 0.5631171095937516, 'J_D_1KI': 0.004610347870460215}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_060.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 115616, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 10.107509136199951, "TIME_S_1KI": 0.0874231000570851, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1385.0806881904603, "W": 104.2, "J_1KI": 11.98000872016382, "W_1KI": 0.9012593412676446, "W_D": 68.209, "J_D": 906.6695648827554, "W_D_1KI": 0.5899615970107944, "J_D_1KI": 0.005102767757151211}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_060.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_060.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 0.027286767959594727}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.8984, 0.0469, 0.3481, ..., 0.4740, 0.2565, 0.1000])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 0.027286767959594727 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '38480', '-m', 'matrices/as-caida_pruned/as-caida_G_060.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 3.4946682453155518}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.1173, 0.8452, 0.0058, ..., 0.9193, 0.4108, 0.2600])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 3.4946682453155518 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '115616', '-m', 'matrices/as-caida_pruned/as-caida_G_060.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_060", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 94180, "MATRIX_DENSITY": 9.564901186217454e-05, "TIME_S": 10.107509136199951}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.3457, 0.6349, 0.8927, ..., 0.7830, 0.7463, 0.9414])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 10.107509136199951 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 94180, 94180, 94180]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=94180, layout=torch.sparse_csr)
tensor([0.3457, 0.6349, 0.8927, ..., 0.7830, 0.7463, 0.9414])
Matrix Type: SuiteSparse
Matrix: as-caida_G_060
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 94180
Density: 9.564901186217454e-05
Time: 10.107509136199951 seconds
[40.73, 39.3, 39.38, 43.93, 40.14, 39.25, 39.56, 39.33, 39.77, 39.89]
[104.2]
13.292520999908447
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 115616, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_060', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 94180, 'MATRIX_DENSITY': 9.564901186217454e-05, 'TIME_S': 10.107509136199951, 'TIME_S_1KI': 0.0874231000570851, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1385.0806881904603, 'W': 104.2}
[40.73, 39.3, 39.38, 43.93, 40.14, 39.25, 39.56, 39.33, 39.77, 39.89, 45.99, 39.65, 39.68, 39.25, 39.25, 39.32, 39.23, 40.28, 39.58, 39.23]
719.82
35.991
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 115616, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_060', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 94180, 'MATRIX_DENSITY': 9.564901186217454e-05, 'TIME_S': 10.107509136199951, 'TIME_S_1KI': 0.0874231000570851, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1385.0806881904603, 'W': 104.2, 'J_1KI': 11.98000872016382, 'W_1KI': 0.9012593412676446, 'W_D': 68.209, 'J_D': 906.6695648827554, 'W_D_1KI': 0.5899615970107944, 'J_D_1KI': 0.005102767757151211}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_065.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 121043, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 10.490303754806519, "TIME_S_1KI": 0.08666592661125815, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1396.8674520850182, "W": 104.75, "J_1KI": 11.540258024710377, "W_1KI": 0.8653949422932347, "W_D": 68.499, "J_D": 913.4512992875575, "W_D_1KI": 0.5659063308080599, "J_D_1KI": 0.004675250372248373}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_065.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_065.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 0.028383970260620117}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.6987, 0.1536, 0.6933, ..., 0.9556, 0.5512, 0.6559])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 0.028383970260620117 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36992', '-m', 'matrices/as-caida_pruned/as-caida_G_065.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 3.2088987827301025}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.2523, 0.5417, 0.6382, ..., 0.2729, 0.0339, 0.5004])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 3.2088987827301025 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '121043', '-m', 'matrices/as-caida_pruned/as-caida_G_065.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_065", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 95068, "MATRIX_DENSITY": 9.655086281283934e-05, "TIME_S": 10.490303754806519}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.9743, 0.5130, 0.5318, ..., 0.8200, 0.9366, 0.1557])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 10.490303754806519 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 95068, 95068, 95068]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 882]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=95068, layout=torch.sparse_csr)
tensor([0.9743, 0.5130, 0.5318, ..., 0.8200, 0.9366, 0.1557])
Matrix Type: SuiteSparse
Matrix: as-caida_G_065
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 95068
Density: 9.655086281283934e-05
Time: 10.490303754806519 seconds
[41.45, 39.82, 40.25, 39.93, 39.31, 39.37, 39.34, 39.68, 44.83, 39.67]
[104.75]
13.33525013923645
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 121043, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_065', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 95068, 'MATRIX_DENSITY': 9.655086281283934e-05, 'TIME_S': 10.490303754806519, 'TIME_S_1KI': 0.08666592661125815, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1396.8674520850182, 'W': 104.75}
[41.45, 39.82, 40.25, 39.93, 39.31, 39.37, 39.34, 39.68, 44.83, 39.67, 40.18, 39.27, 39.88, 39.63, 40.85, 43.9, 39.97, 39.48, 39.27, 39.18]
725.02
36.251
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 121043, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_065', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 95068, 'MATRIX_DENSITY': 9.655086281283934e-05, 'TIME_S': 10.490303754806519, 'TIME_S_1KI': 0.08666592661125815, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1396.8674520850182, 'W': 104.75, 'J_1KI': 11.540258024710377, 'W_1KI': 0.8653949422932347, 'W_D': 68.499, 'J_D': 913.4512992875575, 'W_D_1KI': 0.5659063308080599, 'J_D_1KI': 0.004675250372248373}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_070.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120877, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 10.189666986465454, "TIME_S_1KI": 0.08429781502242324, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1361.2093210601809, "W": 104.18000000000002, "J_1KI": 11.261111055537288, "W_1KI": 0.8618678491358986, "W_D": 68.58550000000002, "J_D": 896.1338250103, "W_D_1KI": 0.567399091638608, "J_D_1KI": 0.004694020298639179}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_070.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_070.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 0.029041528701782227}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.8996, 0.1819, 0.9787, ..., 0.8060, 0.2127, 0.8007])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 0.029041528701782227 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36155', '-m', 'matrices/as-caida_pruned/as-caida_G_070.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 3.14058518409729}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.7195, 0.1055, 0.1671, ..., 0.2379, 0.6021, 0.6792])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 3.14058518409729 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120877', '-m', 'matrices/as-caida_pruned/as-caida_G_070.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_070", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 78684, "MATRIX_DENSITY": 7.991130653390679e-05, "TIME_S": 10.189666986465454}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.8991, 0.4291, 0.0091, ..., 0.3125, 0.3059, 0.1879])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 10.189666986465454 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 78684, 78684, 78684]),
col_indices=tensor([ 106, 329, 1040, ..., 16263, 2242, 2242]),
values=tensor([1., 1., 1., ..., 3., 1., 1.]), size=(31379, 31379),
nnz=78684, layout=torch.sparse_csr)
tensor([0.8991, 0.4291, 0.0091, ..., 0.3125, 0.3059, 0.1879])
Matrix Type: SuiteSparse
Matrix: as-caida_G_070
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 78684
Density: 7.991130653390679e-05
Time: 10.189666986465454 seconds
[40.29, 39.27, 39.66, 39.17, 39.24, 39.24, 39.26, 39.74, 39.48, 39.37]
[104.18]
13.065937042236328
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120877, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_070', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 78684, 'MATRIX_DENSITY': 7.991130653390679e-05, 'TIME_S': 10.189666986465454, 'TIME_S_1KI': 0.08429781502242324, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1361.2093210601809, 'W': 104.18000000000002}
[40.29, 39.27, 39.66, 39.17, 39.24, 39.24, 39.26, 39.74, 39.48, 39.37, 40.49, 39.82, 39.66, 39.31, 39.92, 39.38, 40.13, 39.49, 39.46, 39.17]
711.8900000000001
35.594500000000004
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120877, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_070', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 78684, 'MATRIX_DENSITY': 7.991130653390679e-05, 'TIME_S': 10.189666986465454, 'TIME_S_1KI': 0.08429781502242324, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1361.2093210601809, 'W': 104.18000000000002, 'J_1KI': 11.261111055537288, 'W_1KI': 0.8618678491358986, 'W_D': 68.58550000000002, 'J_D': 896.1338250103, 'W_D_1KI': 0.567399091638608, 'J_D_1KI': 0.004694020298639179}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_075.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120423, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 10.76958441734314, "TIME_S_1KI": 0.08943129150862493, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1443.2409217905997, "W": 104.71, "J_1KI": 11.984761397661574, "W_1KI": 0.8695182813914285, "W_D": 68.72225, "J_D": 947.213861498654, "W_D_1KI": 0.5706737915514478, "J_D_1KI": 0.0047389102708905095}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_075.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_075.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 0.026674270629882812}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.6187, 0.1169, 0.0618, ..., 0.0036, 0.3565, 0.7239])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 0.026674270629882812 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '39363', '-m', 'matrices/as-caida_pruned/as-caida_G_075.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 3.4321558475494385}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.5402, 0.1074, 0.0917, ..., 0.8328, 0.8213, 0.8141])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 3.4321558475494385 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120423', '-m', 'matrices/as-caida_pruned/as-caida_G_075.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_075", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 97492, "MATRIX_DENSITY": 9.901267216465406e-05, "TIME_S": 10.76958441734314}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.9766, 0.5800, 0.0793, ..., 0.9152, 0.2119, 0.8249])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 10.76958441734314 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 97491, 97491, 97492]),
col_indices=tensor([22754, 22754, 106, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=97492, layout=torch.sparse_csr)
tensor([0.9766, 0.5800, 0.0793, ..., 0.9152, 0.2119, 0.8249])
Matrix Type: SuiteSparse
Matrix: as-caida_G_075
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 97492
Density: 9.901267216465406e-05
Time: 10.76958441734314 seconds
[40.44, 39.88, 39.92, 39.3, 39.43, 39.42, 39.28, 39.2, 40.68, 39.23]
[104.71]
13.783219575881958
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120423, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_075', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 97492, 'MATRIX_DENSITY': 9.901267216465406e-05, 'TIME_S': 10.76958441734314, 'TIME_S_1KI': 0.08943129150862493, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1443.2409217905997, 'W': 104.71}
[40.44, 39.88, 39.92, 39.3, 39.43, 39.42, 39.28, 39.2, 40.68, 39.23, 53.53, 39.71, 40.17, 39.13, 39.61, 39.81, 39.45, 39.24, 39.21, 39.43]
719.7549999999999
35.98774999999999
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120423, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_075', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 97492, 'MATRIX_DENSITY': 9.901267216465406e-05, 'TIME_S': 10.76958441734314, 'TIME_S_1KI': 0.08943129150862493, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1443.2409217905997, 'W': 104.71, 'J_1KI': 11.984761397661574, 'W_1KI': 0.8695182813914285, 'W_D': 68.72225, 'J_D': 947.213861498654, 'W_D_1KI': 0.5706737915514478, 'J_D_1KI': 0.0047389102708905095}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_080.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 114354, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 11.189604997634888, "TIME_S_1KI": 0.09785057800894492, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1410.8444130945206, "W": 104.18, "J_1KI": 12.337516948200506, "W_1KI": 0.9110306591811393, "W_D": 68.22325000000001, "J_D": 923.9046948133112, "W_D_1KI": 0.5965969708099411, "J_D_1KI": 0.0052171062735885156}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_080.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_080.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 0.02819371223449707}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.5811, 0.0590, 0.8955, ..., 0.6743, 0.3795, 0.4859])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 0.02819371223449707 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '37242', '-m', 'matrices/as-caida_pruned/as-caida_G_080.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 3.4195492267608643}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.3591, 0.6873, 0.8611, ..., 0.6830, 0.4965, 0.4869])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 3.4195492267608643 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '114354', '-m', 'matrices/as-caida_pruned/as-caida_G_080.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_080", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 98112, "MATRIX_DENSITY": 9.964234287345156e-05, "TIME_S": 11.189604997634888}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.3030, 0.2008, 0.0144, ..., 0.0164, 0.8222, 0.8093])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 11.189604997634888 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 1, 2, ..., 98111, 98111, 98112]),
col_indices=tensor([22754, 22754, 106, ..., 4133, 31329, 12170]),
values=tensor([1., 1., 1., ..., 3., 3., 1.]), size=(31379, 31379),
nnz=98112, layout=torch.sparse_csr)
tensor([0.3030, 0.2008, 0.0144, ..., 0.0164, 0.8222, 0.8093])
Matrix Type: SuiteSparse
Matrix: as-caida_G_080
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 98112
Density: 9.964234287345156e-05
Time: 11.189604997634888 seconds
[40.04, 40.51, 39.79, 39.26, 39.95, 39.3, 39.37, 39.22, 39.38, 44.29]
[104.18]
13.542372941970825
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 114354, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_080', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 98112, 'MATRIX_DENSITY': 9.964234287345156e-05, 'TIME_S': 11.189604997634888, 'TIME_S_1KI': 0.09785057800894492, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1410.8444130945206, 'W': 104.18}
[40.04, 40.51, 39.79, 39.26, 39.95, 39.3, 39.37, 39.22, 39.38, 44.29, 40.85, 39.16, 39.24, 39.16, 39.19, 44.57, 39.27, 39.94, 39.71, 39.05]
719.135
35.95675
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 114354, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_080', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 98112, 'MATRIX_DENSITY': 9.964234287345156e-05, 'TIME_S': 11.189604997634888, 'TIME_S_1KI': 0.09785057800894492, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1410.8444130945206, 'W': 104.18, 'J_1KI': 12.337516948200506, 'W_1KI': 0.9110306591811393, 'W_D': 68.22325000000001, 'J_D': 923.9046948133112, 'W_D_1KI': 0.5965969708099411, 'J_D_1KI': 0.0052171062735885156}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_085.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 123818, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 12.64590573310852, "TIME_S_1KI": 0.10213301566095818, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1404.5768914031983, "W": 105.84, "J_1KI": 11.343882887812743, "W_1KI": 0.8548030173318903, "W_D": 69.798, "J_D": 926.2722776470184, "W_D_1KI": 0.5637144841622381, "J_D_1KI": 0.004552766836503886}

85
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_085.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_085.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 0.027594804763793945}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.3579, 0.6537, 0.3650, ..., 0.6368, 0.7499, 0.3578])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 0.027594804763793945 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '38050', '-m', 'matrices/as-caida_pruned/as-caida_G_085.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 3.2266926765441895}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.5666, 0.3607, 0.0681, ..., 0.1783, 0.0421, 0.2428])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 3.2266926765441895 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '123818', '-m', 'matrices/as-caida_pruned/as-caida_G_085.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_085", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 99166, "MATRIX_DENSITY": 0.0001007127830784073, "TIME_S": 12.64590573310852}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.9239, 0.0838, 0.6171, ..., 0.0890, 0.6862, 0.2789])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 12.64590573310852 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 99165, 99165, 99166]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=99166, layout=torch.sparse_csr)
tensor([0.9239, 0.0838, 0.6171, ..., 0.0890, 0.6862, 0.2789])
Matrix Type: SuiteSparse
Matrix: as-caida_G_085
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 99166
Density: 0.0001007127830784073
Time: 12.64590573310852 seconds
[40.09, 39.23, 39.46, 39.19, 39.83, 39.63, 39.26, 44.52, 39.29, 39.12]
[105.84]
13.270756721496582
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 123818, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_085', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 99166, 'MATRIX_DENSITY': 0.0001007127830784073, 'TIME_S': 12.64590573310852, 'TIME_S_1KI': 0.10213301566095818, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1404.5768914031983, 'W': 105.84}
[40.09, 39.23, 39.46, 39.19, 39.83, 39.63, 39.26, 44.52, 39.29, 39.12, 40.73, 39.74, 39.65, 39.19, 44.6, 39.27, 39.31, 39.32, 39.56, 39.64]
720.8399999999999
36.041999999999994
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 123818, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_085', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 99166, 'MATRIX_DENSITY': 0.0001007127830784073, 'TIME_S': 12.64590573310852, 'TIME_S_1KI': 0.10213301566095818, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1404.5768914031983, 'W': 105.84, 'J_1KI': 11.343882887812743, 'W_1KI': 0.8548030173318903, 'W_D': 69.798, 'J_D': 926.2722776470184, 'W_D_1KI': 0.5637144841622381, 'J_D_1KI': 0.004552766836503886}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_090.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 129633, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 11.644664525985718, "TIME_S_1KI": 0.08982793367418573, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1514.9775815963744, "W": 105.77999999999999, "J_1KI": 11.686666061854423, "W_1KI": 0.8159959269630417, "W_D": 70.1225, "J_D": 1004.2920728445054, "W_D_1KI": 0.5409309357956694, "J_D_1KI": 0.004172787297953989}

89
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_090.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_090.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 0.028497934341430664}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.7120, 0.8564, 0.6416, ..., 0.1452, 0.7450, 0.8345])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 0.028497934341430664 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36844', '-m', 'matrices/as-caida_pruned/as-caida_G_090.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 2.984281539916992}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.2139, 0.0345, 0.3504, ..., 0.9548, 0.7408, 0.1286])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 2.984281539916992 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '129633', '-m', 'matrices/as-caida_pruned/as-caida_G_090.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_090", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 100924, "MATRIX_DENSITY": 0.00010249820421722343, "TIME_S": 11.644664525985718}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.8962, 0.9859, 0.6697, ..., 0.6735, 0.2074, 0.6302])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 11.644664525985718 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 100923, 100923,
100924]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 31211, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=100924, layout=torch.sparse_csr)
tensor([0.8962, 0.9859, 0.6697, ..., 0.6735, 0.2074, 0.6302])
Matrix Type: SuiteSparse
Matrix: as-caida_G_090
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 100924
Density: 0.00010249820421722343
Time: 11.644664525985718 seconds
[40.36, 39.19, 39.41, 39.52, 39.41, 40.55, 39.53, 39.22, 41.03, 39.36]
[105.78]
14.321966171264648
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 129633, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_090', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 100924, 'MATRIX_DENSITY': 0.00010249820421722343, 'TIME_S': 11.644664525985718, 'TIME_S_1KI': 0.08982793367418573, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1514.9775815963744, 'W': 105.77999999999999}
[40.36, 39.19, 39.41, 39.52, 39.41, 40.55, 39.53, 39.22, 41.03, 39.36, 40.6, 40.22, 39.22, 39.14, 39.24, 39.39, 39.44, 39.29, 39.39, 39.6]
713.1499999999999
35.65749999999999
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 129633, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_090', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 100924, 'MATRIX_DENSITY': 0.00010249820421722343, 'TIME_S': 11.644664525985718, 'TIME_S_1KI': 0.08982793367418573, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1514.9775815963744, 'W': 105.77999999999999, 'J_1KI': 11.686666061854423, 'W_1KI': 0.8159959269630417, 'W_D': 70.1225, 'J_D': 1004.2920728445054, 'W_D_1KI': 0.5409309357956694, 'J_D_1KI': 0.004172787297953989}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_095.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 123367, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 10.797792911529541, "TIME_S_1KI": 0.08752578008324383, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1381.98059548378, "W": 106.36000000000001, "J_1KI": 11.202190176333865, "W_1KI": 0.8621430366305415, "W_D": 70.495, "J_D": 915.9714373695851, "W_D_1KI": 0.5714250974733924, "J_D_1KI": 0.004631912079189673}

110
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_095.output Normal file
View File

@ -0,0 +1,110 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_095.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 0.02869391441345215}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.6977, 0.1794, 0.1733, ..., 0.5021, 0.2371, 0.5915])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 0.02869391441345215 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36593', '-m', 'matrices/as-caida_pruned/as-caida_G_095.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 3.28193736076355}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.2038, 0.2117, 0.0495, ..., 0.7353, 0.4058, 0.7220])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 3.28193736076355 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '117073', '-m', 'matrices/as-caida_pruned/as-caida_G_095.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 9.964244604110718}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.2057, 0.4266, 0.2763, ..., 0.1377, 0.7100, 0.3501])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 9.964244604110718 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '123367', '-m', 'matrices/as-caida_pruned/as-caida_G_095.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_095", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102290, "MATRIX_DENSITY": 0.00010388551097241275, "TIME_S": 10.797792911529541}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.2200, 0.4810, 0.3293, ..., 0.7198, 0.3850, 0.3915])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 10.797792911529541 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102289, 102289,
102290]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 25970, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102290, layout=torch.sparse_csr)
tensor([0.2200, 0.4810, 0.3293, ..., 0.7198, 0.3850, 0.3915])
Matrix Type: SuiteSparse
Matrix: as-caida_G_095
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102290
Density: 0.00010388551097241275
Time: 10.797792911529541 seconds
[40.51, 39.95, 39.69, 39.82, 39.8, 39.17, 39.45, 39.15, 39.47, 39.6]
[106.36]
12.9934241771698
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 123367, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_095', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102290, 'MATRIX_DENSITY': 0.00010388551097241275, 'TIME_S': 10.797792911529541, 'TIME_S_1KI': 0.08752578008324383, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1381.98059548378, 'W': 106.36000000000001}
[40.51, 39.95, 39.69, 39.82, 39.8, 39.17, 39.45, 39.15, 39.47, 39.6, 41.41, 39.21, 39.7, 39.75, 39.19, 39.65, 39.19, 40.91, 42.9, 39.08]
717.3000000000001
35.865
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 123367, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_095', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102290, 'MATRIX_DENSITY': 0.00010388551097241275, 'TIME_S': 10.797792911529541, 'TIME_S_1KI': 0.08752578008324383, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1381.98059548378, 'W': 106.36000000000001, 'J_1KI': 11.202190176333865, 'W_1KI': 0.8621430366305415, 'W_D': 70.495, 'J_D': 915.9714373695851, 'W_D_1KI': 0.5714250974733924, 'J_D_1KI': 0.004631912079189673}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_100.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120963, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 11.165674448013306, "TIME_S_1KI": 0.09230652718610903, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1373.6955507874488, "W": 105.85, "J_1KI": 11.356328387915717, "W_1KI": 0.8750609690566536, "W_D": 69.60775, "J_D": 903.3524466256499, "W_D_1KI": 0.5754466241743342, "J_D_1KI": 0.004757211909214671}

89
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_100.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_100.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 0.027922391891479492}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.7044, 0.9718, 0.4063, ..., 0.8134, 0.1834, 0.4981])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 0.027922391891479492 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '37604', '-m', 'matrices/as-caida_pruned/as-caida_G_100.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 3.264139175415039}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.5057, 0.0011, 0.7436, ..., 0.8079, 0.3676, 0.2944])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 3.264139175415039 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120963', '-m', 'matrices/as-caida_pruned/as-caida_G_100.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_100", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 102888, "MATRIX_DENSITY": 0.00010449283852702711, "TIME_S": 11.165674448013306}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.8382, 0.6727, 0.9499, ..., 0.3590, 0.1899, 0.3958])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 11.165674448013306 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 102886, 102887,
102888]),
col_indices=tensor([ 106, 329, 1040, ..., 25970, 5128, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=102888, layout=torch.sparse_csr)
tensor([0.8382, 0.6727, 0.9499, ..., 0.3590, 0.1899, 0.3958])
Matrix Type: SuiteSparse
Matrix: as-caida_G_100
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 102888
Density: 0.00010449283852702711
Time: 11.165674448013306 seconds
[41.25, 40.42, 39.99, 40.1, 40.02, 39.82, 39.84, 39.36, 44.69, 39.26]
[105.85]
12.97775673866272
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120963, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_100', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102888, 'MATRIX_DENSITY': 0.00010449283852702711, 'TIME_S': 11.165674448013306, 'TIME_S_1KI': 0.09230652718610903, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1373.6955507874488, 'W': 105.85}
[41.25, 40.42, 39.99, 40.1, 40.02, 39.82, 39.84, 39.36, 44.69, 39.26, 40.13, 39.22, 39.32, 39.59, 39.64, 44.99, 39.36, 39.16, 39.44, 39.13]
724.845
36.24225
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120963, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_100', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 102888, 'MATRIX_DENSITY': 0.00010449283852702711, 'TIME_S': 11.165674448013306, 'TIME_S_1KI': 0.09230652718610903, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1373.6955507874488, 'W': 105.85, 'J_1KI': 11.356328387915717, 'W_1KI': 0.8750609690566536, 'W_D': 69.60775, 'J_D': 903.3524466256499, 'W_D_1KI': 0.5754466241743342, 'J_D_1KI': 0.004757211909214671}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_105.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 130315, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 11.23905086517334, "TIME_S_1KI": 0.0862452585287445, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1479.8992812085153, "W": 106.81, "J_1KI": 11.356323379568856, "W_1KI": 0.8196293596285923, "W_D": 71.0115, "J_D": 983.8954012502431, "W_D_1KI": 0.5449219199631662, "J_D_1KI": 0.004181574799241578}

110
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_105.output Normal file
View File

@ -0,0 +1,110 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_105.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 0.02874302864074707}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.0653, 0.6055, 0.4558, ..., 0.6884, 0.5872, 0.3001])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 0.02874302864074707 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36530', '-m', 'matrices/as-caida_pruned/as-caida_G_105.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 3.1528003215789795}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.0149, 0.2762, 0.5141, ..., 0.7208, 0.6938, 0.9073])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 3.1528003215789795 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '121658', '-m', 'matrices/as-caida_pruned/as-caida_G_105.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 9.802452087402344}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.9575, 0.4028, 0.3578, ..., 0.2947, 0.5779, 0.9432])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 9.802452087402344 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '130315', '-m', 'matrices/as-caida_pruned/as-caida_G_105.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_105", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104726, "MATRIX_DENSITY": 0.00010635950749923647, "TIME_S": 11.23905086517334}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.3162, 0.7565, 0.7515, ..., 0.4628, 0.3648, 0.1630])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 11.23905086517334 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104725, 104725,
104726]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104726, layout=torch.sparse_csr)
tensor([0.3162, 0.7565, 0.7515, ..., 0.4628, 0.3648, 0.1630])
Matrix Type: SuiteSparse
Matrix: as-caida_G_105
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104726
Density: 0.00010635950749923647
Time: 11.23905086517334 seconds
[40.79, 41.78, 39.38, 39.23, 39.79, 39.25, 39.89, 39.93, 39.66, 39.98]
[106.81]
13.855437517166138
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 130315, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_105', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104726, 'MATRIX_DENSITY': 0.00010635950749923647, 'TIME_S': 11.23905086517334, 'TIME_S_1KI': 0.0862452585287445, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1479.8992812085153, 'W': 106.81}
[40.79, 41.78, 39.38, 39.23, 39.79, 39.25, 39.89, 39.93, 39.66, 39.98, 41.03, 39.28, 39.81, 39.38, 39.32, 39.19, 40.5, 39.21, 39.67, 39.6]
715.97
35.798500000000004
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 130315, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_105', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104726, 'MATRIX_DENSITY': 0.00010635950749923647, 'TIME_S': 11.23905086517334, 'TIME_S_1KI': 0.0862452585287445, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1479.8992812085153, 'W': 106.81, 'J_1KI': 11.356323379568856, 'W_1KI': 0.8196293596285923, 'W_D': 71.0115, 'J_D': 983.8954012502431, 'W_D_1KI': 0.5449219199631662, 'J_D_1KI': 0.004181574799241578}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_110.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 120860, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 10.835078716278076, "TIME_S_1KI": 0.08964983217175307, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1468.0531423473358, "W": 105.19, "J_1KI": 12.146724659501372, "W_1KI": 0.8703458547079265, "W_D": 69.35275, "J_D": 967.9011556985379, "W_D_1KI": 0.5738271553863975, "J_D_1KI": 0.004747866584365361}

89
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_110.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_110.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 0.033330678939819336}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.1609, 0.2446, 0.3388, ..., 0.7294, 0.1175, 0.4322])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 0.033330678939819336 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '31502', '-m', 'matrices/as-caida_pruned/as-caida_G_110.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 2.7367961406707764}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.5708, 0.9990, 0.7261, ..., 0.2817, 0.6295, 0.6453])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 2.7367961406707764 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '120860', '-m', 'matrices/as-caida_pruned/as-caida_G_110.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_110", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 104846, "MATRIX_DENSITY": 0.0001064813792493263, "TIME_S": 10.835078716278076}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.6634, 0.1694, 0.2296, ..., 0.0338, 0.8523, 0.3205])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 10.835078716278076 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 104844, 104844,
104846]),
col_indices=tensor([ 106, 329, 1040, ..., 882, 2616, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=104846, layout=torch.sparse_csr)
tensor([0.6634, 0.1694, 0.2296, ..., 0.0338, 0.8523, 0.3205])
Matrix Type: SuiteSparse
Matrix: as-caida_G_110
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 104846
Density: 0.0001064813792493263
Time: 10.835078716278076 seconds
[39.93, 39.96, 39.67, 39.83, 41.04, 39.69, 39.85, 39.54, 39.73, 39.63]
[105.19]
13.956204414367676
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120860, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_110', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104846, 'MATRIX_DENSITY': 0.0001064813792493263, 'TIME_S': 10.835078716278076, 'TIME_S_1KI': 0.08964983217175307, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1468.0531423473358, 'W': 105.19}
[39.93, 39.96, 39.67, 39.83, 41.04, 39.69, 39.85, 39.54, 39.73, 39.63, 41.61, 39.22, 39.41, 39.14, 39.63, 39.34, 39.68, 41.03, 39.64, 39.52]
716.7449999999999
35.83725
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 120860, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_110', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 104846, 'MATRIX_DENSITY': 0.0001064813792493263, 'TIME_S': 10.835078716278076, 'TIME_S_1KI': 0.08964983217175307, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1468.0531423473358, 'W': 105.19, 'J_1KI': 12.146724659501372, 'W_1KI': 0.8703458547079265, 'W_D': 69.35275, 'J_D': 967.9011556985379, 'W_D_1KI': 0.5738271553863975, 'J_D_1KI': 0.004747866584365361}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_115.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 122384, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 10.623056411743164, "TIME_S_1KI": 0.08680102310549716, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1373.5208738970757, "W": 106.31, "J_1KI": 11.223042831555397, "W_1KI": 0.8686593018695254, "W_D": 70.50425, "J_D": 910.9120409505963, "W_D_1KI": 0.5760904203163812, "J_D_1KI": 0.004707236406036584}

89
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_115.output Normal file
View File

@ -0,0 +1,89 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_115.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 0.029188871383666992}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.1969, 0.7804, 0.2686, ..., 0.5827, 0.0798, 0.7838])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 0.029188871383666992 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '35972', '-m', 'matrices/as-caida_pruned/as-caida_G_115.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 3.086211681365967}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.5660, 0.8309, 0.9656, ..., 0.1156, 0.8355, 0.1569])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 3.086211681365967 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '122384', '-m', 'matrices/as-caida_pruned/as-caida_G_115.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_115", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106312, "MATRIX_DENSITY": 0.00010797024579625715, "TIME_S": 10.623056411743164}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.4590, 0.4390, 0.2483, ..., 0.8018, 0.3092, 0.2454])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 10.623056411743164 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106311, 106311,
106312]),
col_indices=tensor([ 106, 329, 1040, ..., 160, 882, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106312, layout=torch.sparse_csr)
tensor([0.4590, 0.4390, 0.2483, ..., 0.8018, 0.3092, 0.2454])
Matrix Type: SuiteSparse
Matrix: as-caida_G_115
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106312
Density: 0.00010797024579625715
Time: 10.623056411743164 seconds
[39.96, 41.64, 39.36, 39.3, 39.65, 39.66, 39.8, 40.15, 39.67, 40.32]
[106.31]
12.919959306716919
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122384, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_115', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106312, 'MATRIX_DENSITY': 0.00010797024579625715, 'TIME_S': 10.623056411743164, 'TIME_S_1KI': 0.08680102310549716, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1373.5208738970757, 'W': 106.31}
[39.96, 41.64, 39.36, 39.3, 39.65, 39.66, 39.8, 40.15, 39.67, 40.32, 39.97, 39.29, 39.38, 39.24, 39.46, 39.22, 39.78, 41.14, 39.67, 39.16]
716.115
35.80575
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 122384, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_115', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106312, 'MATRIX_DENSITY': 0.00010797024579625715, 'TIME_S': 10.623056411743164, 'TIME_S_1KI': 0.08680102310549716, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1373.5208738970757, 'W': 106.31, 'J_1KI': 11.223042831555397, 'W_1KI': 0.8686593018695254, 'W_D': 70.50425, 'J_D': 910.9120409505963, 'W_D_1KI': 0.5760904203163812, 'J_D_1KI': 0.004707236406036584}

1
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_120.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 121667, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 10.714912414550781, "TIME_S_1KI": 0.08806753198937083, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1471.0119538068773, "W": 105.95, "J_1KI": 12.09047608477958, "W_1KI": 0.8708195320012823, "W_D": 70.13625000000002, "J_D": 973.7731207662823, "W_D_1KI": 0.5764607494226045, "J_D_1KI": 0.004738020576019828}

110
pytorch/output_as-caida_16core/epyc_7313p_16_csr_10_10_10_as-caida_G_120.output Normal file
View File

@ -0,0 +1,110 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_120.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 0.02883291244506836}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.3732, 0.0574, 0.5467, ..., 0.3088, 0.7435, 0.5865])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 0.02883291244506836 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '36416', '-m', 'matrices/as-caida_pruned/as-caida_G_120.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 3.683297872543335}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.3740, 0.4220, 0.4478, ..., 0.4704, 0.3788, 0.3735])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 3.683297872543335 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '103811', '-m', 'matrices/as-caida_pruned/as-caida_G_120.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 8.958942651748657}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.8035, 0.5449, 0.4888, ..., 0.6708, 0.9679, 0.8623])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 8.958942651748657 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '121667', '-m', 'matrices/as-caida_pruned/as-caida_G_120.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_120", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 106510, "MATRIX_DENSITY": 0.0001081713341839054, "TIME_S": 10.714912414550781}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.3594, 0.3055, 0.9207, ..., 0.9236, 0.8473, 0.9639])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 10.714912414550781 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 0, 0, ..., 106509, 106509,
106510]),
col_indices=tensor([ 106, 329, 1040, ..., 155, 160, 12170]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=106510, layout=torch.sparse_csr)
tensor([0.3594, 0.3055, 0.9207, ..., 0.9236, 0.8473, 0.9639])
Matrix Type: SuiteSparse
Matrix: as-caida_G_120
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 106510
Density: 0.0001081713341839054
Time: 10.714912414550781 seconds
[40.81, 39.31, 39.39, 39.28, 39.62, 39.64, 39.64, 39.98, 39.31, 39.76]
[105.95]
13.884020328521729
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 121667, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_120', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106510, 'MATRIX_DENSITY': 0.0001081713341839054, 'TIME_S': 10.714912414550781, 'TIME_S_1KI': 0.08806753198937083, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1471.0119538068773, 'W': 105.95}
[40.81, 39.31, 39.39, 39.28, 39.62, 39.64, 39.64, 39.98, 39.31, 39.76, 40.19, 39.69, 39.9, 40.2, 39.34, 39.26, 39.38, 39.89, 42.24, 39.65]
716.2749999999999
35.81374999999999
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 121667, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_120', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 106510, 'MATRIX_DENSITY': 0.0001081713341839054, 'TIME_S': 10.714912414550781, 'TIME_S_1KI': 0.08806753198937083, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1471.0119538068773, 'W': 105.95, 'J_1KI': 12.09047608477958, 'W_1KI': 0.8708195320012823, 'W_D': 70.13625000000002, 'J_D': 973.7731207662823, 'W_D_1KI': 0.5764607494226045, 'J_D_1KI': 0.004738020576019828}

1
pytorch/output_as-caida_16core/xeon_4216_16_csr_10_10_10_as-caida_G_005.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 118951, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 10.958328008651733, "TIME_S_1KI": 0.0921247236984282, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1244.7510617017745, "W": 85.66, "J_1KI": 10.46440182681755, "W_1KI": 0.7201284562550967, "W_D": 68.79275, "J_D": 999.64801073879, "W_D_1KI": 0.578328471387378, "J_D_1KI": 0.004861905081818379}

85
pytorch/output_as-caida_16core/xeon_4216_16_csr_10_10_10_as-caida_G_005.output Normal file
View File

@ -0,0 +1,85 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 0.02218484878540039}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.8422, 0.7050, 0.1082, ..., 0.1730, 0.8671, 0.7264])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 0.02218484878540039 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '47329', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 4.177785634994507}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.4803, 0.7198, 0.6367, ..., 0.1841, 0.4829, 0.5995])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 4.177785634994507 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '118951', '-m', 'matrices/as-caida_pruned/as-caida_G_005.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_005", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 70026, "MATRIX_DENSITY": 7.111825976492498e-05, "TIME_S": 10.958328008651733}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.4882, 0.5169, 0.2032, ..., 0.7665, 0.8923, 0.8813])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 10.958328008651733 seconds
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If you miss a functionality in the sparse tensor support, please submit a feature request to https://github.com/pytorch/pytorch/issues. (Triggered internally at ../aten/src/ATen/SparseCsrTensorImpl.cpp:53.)
matrix = matrix.to_sparse_csr().type(torch.float32)
tensor(crow_indices=tensor([ 0, 63, 63, ..., 70025, 70025, 70026]),
col_indices=tensor([ 111, 761, 822, ..., 978, 978, 12170]),
values=tensor([4., 3., 3., ..., 1., 1., 1.]), size=(31379, 31379),
nnz=70026, layout=torch.sparse_csr)
tensor([0.4882, 0.5169, 0.2032, ..., 0.7665, 0.8923, 0.8813])
Matrix Type: SuiteSparse
Matrix: as-caida_G_005
Matrix Format: csr
Shape: torch.Size([31379, 31379])
Rows: 31379
Size: 984641641
NNZ: 70026
Density: 7.111825976492498e-05
Time: 10.958328008651733 seconds
[19.13, 18.64, 19.2, 18.61, 18.66, 18.55, 19.69, 18.7, 18.45, 18.58]
[85.66]
14.531298875808716
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 118951, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 10.958328008651733, 'TIME_S_1KI': 0.0921247236984282, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1244.7510617017745, 'W': 85.66}
[19.13, 18.64, 19.2, 18.61, 18.66, 18.55, 19.69, 18.7, 18.45, 18.58, 18.68, 18.33, 18.75, 18.56, 18.36, 18.54, 18.71, 18.72, 19.33, 18.7]
337.345
16.867250000000002
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 118951, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'as-caida_G_005', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [31379, 31379], 'MATRIX_ROWS': 31379, 'MATRIX_SIZE': 984641641, 'MATRIX_NNZ': 70026, 'MATRIX_DENSITY': 7.111825976492498e-05, 'TIME_S': 10.958328008651733, 'TIME_S_1KI': 0.0921247236984282, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1244.7510617017745, 'W': 85.66, 'J_1KI': 10.46440182681755, 'W_1KI': 0.7201284562550967, 'W_D': 68.79275, 'J_D': 999.64801073879, 'W_D_1KI': 0.578328471387378, 'J_D_1KI': 0.004861905081818379}

1
pytorch/output_as-caida_16core/xeon_4216_16_csr_10_10_10_as-caida_G_010.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 111707, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "as-caida_G_010", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [31379, 31379], "MATRIX_ROWS": 31379, "MATRIX_SIZE": 984641641, "MATRIX_NNZ": 74994, "MATRIX_DENSITY": 7.616375021864427e-05, "TIME_S": 10.419404983520508, "TIME_S_1KI": 0.09327441416849891, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1175.4035894393921, "W": 84.72, "J_1KI": 10.52220173703879, "W_1KI": 0.7584126330489583, "W_D": 67.76225, "J_D": 940.1321043258905, "W_D_1KI": 0.6066070165701344, "J_D_1KI": 0.005430340234453834}

Some files were not shown because too many files have changed in this diff Show More