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

More datagit status!

Browse Source
This commit is contained in:
cephi 2024-12-16 09:46:15 -05:00
parent 1ba4a701cf
commit a1af08726d
490 changed files with 27153 additions and 17002 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) parameter_list = enumerate([(size, density)
for size in args.synthetic_size for size in args.synthetic_size
for density in args.synthetic_density for density in args.synthetic_density
if size ** 2 * density < 10000000]) if size ** 2 * density <= 10000000])
#for i, matrix in enumerate(glob.glob(f'{args.matrix_dir.rstrip("/")}/*.mtx')): #for i, matrix in enumerate(glob.glob(f'{args.matrix_dir.rstrip("/")}/*.mtx')):
for i, parameter in parameter_list: for i, parameter in parameter_list:

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1345, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.307875871658325, "TIME_S_1KI": 7.663848231716227, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 549.730076084137, "W": 38.131256134653135, "J_1KI": 408.7212461592096, "W_1KI": 28.350376308292294, "W_D": 16.28825613465314, "J_D": 234.82426732969293, "W_D_1KI": 12.110227609407538, "J_D_1KI": 9.003886698444267}

71
pytorch/output_389000+_16core/altra_16_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,71 @@
['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 1000 -m matrices/389000+_cols/amazon0312.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 7.806152820587158}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.5486, 0.8485, 0.8195, ..., 0.3778, 0.3275, 0.7623])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 7.806152820587158 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 1345 -m matrices/389000+_cols/amazon0312.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.307875871658325}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6343, 0.9450, 0.3421, ..., 0.5967, 0.9759, 0.2168])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.307875871658325 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6343, 0.9450, 0.3421, ..., 0.5967, 0.9759, 0.2168])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.307875871658325 seconds
[24.12, 24.04, 24.4, 24.52, 24.56, 24.68, 24.68, 24.64, 24.56, 24.28]
[24.24, 24.0, 24.0, 24.8, 25.52, 28.6, 36.84, 43.04, 50.0, 55.48, 57.44, 57.44, 57.68, 57.56]
14.416783809661865
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1345, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.307875871658325, 'TIME_S_1KI': 7.663848231716227, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 549.730076084137, 'W': 38.131256134653135}
[24.12, 24.04, 24.4, 24.52, 24.56, 24.68, 24.68, 24.64, 24.56, 24.28, 24.12, 24.24, 24.0, 23.96, 23.96, 24.08, 24.08, 24.08, 24.08, 24.08]
436.85999999999996
21.842999999999996
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1345, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.307875871658325, 'TIME_S_1KI': 7.663848231716227, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 549.730076084137, 'W': 38.131256134653135, 'J_1KI': 408.7212461592096, 'W_1KI': 28.350376308292294, 'W_D': 16.28825613465314, 'J_D': 234.82426732969293, 'W_D_1KI': 12.110227609407538, 'J_D_1KI': 9.003886698444267}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1641, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.180182933807373, "TIME_S_1KI": 6.203645907256169, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 572.6159508895875, "W": 39.25282662523282, "J_1KI": 348.94329731236286, "W_1KI": 23.920064975766497, "W_D": 17.70882662523282, "J_D": 258.33443012809755, "W_D_1KI": 10.791484841701902, "J_D_1KI": 6.576163827971908}

71
pytorch/output_389000+_16core/altra_16_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,71 @@
['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 1000 -m matrices/389000+_cols/darcy003.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 6.3963303565979}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.2730, 0.2238, 0.6515, ..., 0.6572, 0.5843, 0.9667])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 6.3963303565979 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 1641 -m matrices/389000+_cols/darcy003.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.180182933807373}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.4101, 0.8547, 0.0587, ..., 0.2935, 0.9064, 0.8922])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.180182933807373 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.4101, 0.8547, 0.0587, ..., 0.2935, 0.9064, 0.8922])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.180182933807373 seconds
[23.68, 23.72, 23.96, 23.96, 24.16, 24.28, 24.28, 24.52, 24.76, 24.52]
[24.16, 23.84, 23.72, 26.96, 28.84, 34.04, 41.68, 44.88, 51.8, 55.52, 55.8, 56.2, 56.16, 56.16]
14.587890863418579
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1641, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.180182933807373, 'TIME_S_1KI': 6.203645907256169, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 572.6159508895875, 'W': 39.25282662523282}
[23.68, 23.72, 23.96, 23.96, 24.16, 24.28, 24.28, 24.52, 24.76, 24.52, 24.08, 24.08, 24.12, 24.04, 23.6, 23.44, 23.28, 23.28, 23.4, 23.72]
430.88
21.544
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1641, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.180182933807373, 'TIME_S_1KI': 6.203645907256169, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 572.6159508895875, 'W': 39.25282662523282, 'J_1KI': 348.94329731236286, 'W_1KI': 23.920064975766497, 'W_D': 17.70882662523282, 'J_D': 258.33443012809755, 'W_D_1KI': 10.791484841701902, 'J_D_1KI': 6.576163827971908}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1801, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.381673336029053, "TIME_S_1KI": 5.764393856762384, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 582.8333181858063, "W": 39.805332297033694, "J_1KI": 323.6165009360391, "W_1KI": 22.10179472350566, "W_D": 18.366332297033694, "J_D": 268.92151824545857, "W_D_1KI": 10.197852469202495, "J_D_1KI": 5.66232785630344}

74
pytorch/output_389000+_16core/altra_16_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,74 @@
['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 1000 -m matrices/389000+_cols/helm2d03.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 5.829137325286865}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.8248, 0.9604, 0.9464, ..., 0.7437, 0.8759, 0.5369])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 5.829137325286865 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 1801 -m matrices/389000+_cols/helm2d03.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.381673336029053}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.8790, 0.0885, 0.6163, ..., 0.1605, 0.4532, 0.8862])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.381673336029053 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.8790, 0.0885, 0.6163, ..., 0.1605, 0.4532, 0.8862])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.381673336029053 seconds
[23.72, 23.72, 23.68, 23.68, 23.68, 23.6, 23.68, 23.64, 23.64, 23.72]
[23.88, 24.0, 24.28, 26.28, 27.2, 34.12, 40.64, 47.44, 53.0, 57.04, 57.04, 56.96, 57.52, 57.24]
14.642091512680054
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1801, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.381673336029053, 'TIME_S_1KI': 5.764393856762384, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 582.8333181858063, 'W': 39.805332297033694}
[23.72, 23.72, 23.68, 23.68, 23.68, 23.6, 23.68, 23.64, 23.64, 23.72, 23.68, 24.04, 24.0, 24.12, 24.0, 23.92, 24.08, 24.04, 23.76, 23.88]
428.78000000000003
21.439
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1801, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.381673336029053, 'TIME_S_1KI': 5.764393856762384, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 582.8333181858063, 'W': 39.805332297033694, 'J_1KI': 323.6165009360391, 'W_1KI': 22.10179472350566, 'W_D': 18.366332297033694, 'J_D': 268.92151824545857, 'W_D_1KI': 10.197852469202495, 'J_D_1KI': 5.66232785630344}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 2142, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.489102602005005, "TIME_S_1KI": 4.896873296921104, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 527.8806602096558, "W": 35.901030267154354, "J_1KI": 246.44288525194017, "W_1KI": 16.760518332004835, "W_D": 14.631030267154355, "J_D": 215.1313725399972, "W_D_1KI": 6.830546343209316, "J_D_1KI": 3.1888638390332944}

71
pytorch/output_389000+_16core/altra_16_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,71 @@
['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 1000 -m matrices/389000+_cols/language.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 4.9012720584869385}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.0783, 0.8612, 0.3161, ..., 0.8531, 0.6998, 0.6080])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 4.9012720584869385 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 2142 -m matrices/389000+_cols/language.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.489102602005005}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.5677, 0.4069, 0.3735, ..., 0.4488, 0.2885, 0.1400])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.489102602005005 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.5677, 0.4069, 0.3735, ..., 0.4488, 0.2885, 0.1400])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.489102602005005 seconds
[23.72, 23.8, 23.88, 23.8, 23.8, 23.52, 23.48, 23.36, 23.36, 23.24]
[23.28, 23.4, 23.56, 24.68, 26.6, 30.12, 36.56, 41.44, 45.96, 49.92, 50.84, 50.96, 50.76, 51.04]
14.703774690628052
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 2142, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.489102602005005, 'TIME_S_1KI': 4.896873296921104, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 527.8806602096558, 'W': 35.901030267154354}
[23.72, 23.8, 23.88, 23.8, 23.8, 23.52, 23.48, 23.36, 23.36, 23.24, 23.6, 23.64, 23.64, 23.68, 23.68, 23.68, 23.76, 23.64, 23.64, 23.52]
425.4
21.27
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 2142, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.489102602005005, 'TIME_S_1KI': 4.896873296921104, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 527.8806602096558, 'W': 35.901030267154354, 'J_1KI': 246.44288525194017, 'W_1KI': 16.760518332004835, 'W_D': 14.631030267154355, 'J_D': 215.1313725399972, 'W_D_1KI': 6.830546343209316, 'J_D_1KI': 3.1888638390332944}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 12.712969779968262, "TIME_S_1KI": 12.712969779968262, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 642.9576612091064, "W": 41.13704135154824, "J_1KI": 642.9576612091064, "W_1KI": 41.13704135154824, "W_D": 19.36904135154824, "J_D": 302.73138558578484, "W_D_1KI": 19.36904135154824, "J_D_1KI": 19.36904135154824}

51
pytorch/output_389000+_16core/altra_16_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,51 @@
['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 1000 -m matrices/389000+_cols/marine1.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 12.712969779968262}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.7918, 0.6380, 0.4821, ..., 0.8085, 0.1927, 0.4528])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 12.712969779968262 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.7918, 0.6380, 0.4821, ..., 0.8085, 0.1927, 0.4528])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 12.712969779968262 seconds
[24.16, 24.24, 24.24, 24.08, 24.2, 24.48, 24.48, 24.6, 24.28, 24.24]
[23.96, 23.92, 24.76, 26.04, 30.04, 34.84, 41.8, 41.8, 48.8, 54.68, 57.92, 59.36, 59.6, 60.0, 59.68]
15.629652500152588
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 12.712969779968262, 'TIME_S_1KI': 12.712969779968262, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 642.9576612091064, 'W': 41.13704135154824}
[24.16, 24.24, 24.24, 24.08, 24.2, 24.48, 24.48, 24.6, 24.28, 24.24, 24.24, 24.16, 24.08, 24.12, 24.12, 24.0, 23.92, 24.0, 24.0, 24.08]
435.36
21.768
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 12.712969779968262, 'TIME_S_1KI': 12.712969779968262, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 642.9576612091064, 'W': 41.13704135154824, 'J_1KI': 642.9576612091064, 'W_1KI': 41.13704135154824, 'W_D': 19.36904135154824, 'J_D': 302.73138558578484, 'W_D_1KI': 19.36904135154824, 'J_D_1KI': 19.36904135154824}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1694, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.441875219345093, "TIME_S_1KI": 6.164034958291081, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 563.8270891189575, "W": 38.54107447042048, "J_1KI": 332.8377149462559, "W_1KI": 22.751519758217523, "W_D": 16.675074470420483, "J_D": 243.94386582851405, "W_D_1KI": 9.843609486670887, "J_D_1KI": 5.81086746556723}

93
pytorch/output_389000+_16core/altra_16_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,93 @@
['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 1000 -m matrices/389000+_cols/mario002.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 7.140163421630859}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.1948, 0.1869, 0.5638, ..., 0.6155, 0.6170, 0.7726])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 7.140163421630859 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 1470 -m matrices/389000+_cols/mario002.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 9.107451438903809}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.1872, 0.8693, 0.3135, ..., 0.4431, 0.3648, 0.2379])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 9.107451438903809 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 1694 -m matrices/389000+_cols/mario002.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.441875219345093}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.6416, 0.1879, 0.5321, ..., 0.0693, 0.5314, 0.2281])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.441875219345093 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.6416, 0.1879, 0.5321, ..., 0.0693, 0.5314, 0.2281])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.441875219345093 seconds
[24.04, 23.88, 24.2, 24.4, 24.52, 24.32, 24.64, 24.08, 24.08, 24.2]
[24.04, 23.88, 24.16, 26.0, 26.68, 32.8, 39.08, 45.28, 50.52, 54.72, 54.84, 55.48, 55.4, 55.76]
14.629251956939697
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1694, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.441875219345093, 'TIME_S_1KI': 6.164034958291081, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 563.8270891189575, 'W': 38.54107447042048}
[24.04, 23.88, 24.2, 24.4, 24.52, 24.32, 24.64, 24.08, 24.08, 24.2, 24.0, 23.96, 24.08, 24.08, 24.44, 24.72, 24.6, 24.56, 24.48, 24.32]
437.32
21.866
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1694, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.441875219345093, 'TIME_S_1KI': 6.164034958291081, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 563.8270891189575, 'W': 38.54107447042048, 'J_1KI': 332.8377149462559, 'W_1KI': 22.751519758217523, 'W_D': 16.675074470420483, 'J_D': 243.94386582851405, 'W_D_1KI': 9.843609486670887, 'J_D_1KI': 5.81086746556723}

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 33.385778188705444, "TIME_S_1KI": 33.385778188705444, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1758.8809212875371, "W": 47.990539203077404, "J_1KI": 1758.8809212875371, "W_1KI": 47.990539203077404, "W_D": 26.191539203077404, "J_D": 959.9350073668963, "W_D_1KI": 26.191539203077404, "J_D_1KI": 26.191539203077404}

51
pytorch/output_389000+_16core/altra_16_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,51 @@
['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 1000 -m matrices/389000+_cols/test1.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 33.385778188705444}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3716, 0.7020, 0.0579, ..., 0.5562, 0.4218, 0.2724])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 33.385778188705444 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3716, 0.7020, 0.0579, ..., 0.5562, 0.4218, 0.2724])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 33.385778188705444 seconds
[24.12, 24.32, 24.32, 24.28, 24.08, 24.2, 24.0, 24.0, 24.08, 24.08]
[24.24, 24.16, 24.36, 28.92, 31.16, 35.28, 37.04, 40.44, 46.16, 49.16, 54.2, 56.72, 56.48, 56.72, 56.72, 56.48, 57.24, 56.52, 56.44, 56.56, 56.52, 57.2, 56.84, 56.8, 56.92, 56.92, 58.08, 57.8, 57.88, 57.44, 57.4, 56.88, 56.44, 57.16, 57.12]
36.65057635307312
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 33.385778188705444, 'TIME_S_1KI': 33.385778188705444, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1758.8809212875371, 'W': 47.990539203077404}
[24.12, 24.32, 24.32, 24.28, 24.08, 24.2, 24.0, 24.0, 24.08, 24.08, 24.24, 24.24, 24.24, 24.28, 24.48, 24.36, 24.4, 24.16, 24.16, 24.32]
435.98
21.799
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 33.385778188705444, 'TIME_S_1KI': 33.385778188705444, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1758.8809212875371, 'W': 47.990539203077404, 'J_1KI': 1758.8809212875371, 'W_1KI': 47.990539203077404, 'W_D': 26.191539203077404, 'J_D': 959.9350073668963, 'W_D_1KI': 26.191539203077404, 'J_D_1KI': 26.191539203077404}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 20402, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.640971422195435, "TIME_S_1KI": 0.5215651123515065, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1986.5667917442322, "W": 146.96, "J_1KI": 97.3711788914926, "W_1KI": 7.203215371042055, "W_D": 111.28150000000001, "J_D": 1504.2741728054286, "W_D_1KI": 5.454440741103813, "J_D_1KI": 0.2673483355114113}

93
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,93 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 0.5630712509155273}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.4842, 0.5105, 0.4860, ..., 0.7675, 0.4934, 0.1706])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 0.5630712509155273 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '18647', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 9.59645700454712}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6749, 0.4854, 0.2428, ..., 0.8655, 0.6324, 0.8376])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 9.59645700454712 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '20402', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.640971422195435}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.0724, 0.4329, 0.4595, ..., 0.8349, 0.8167, 0.6766])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.640971422195435 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.0724, 0.4329, 0.4595, ..., 0.8349, 0.8167, 0.6766])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.640971422195435 seconds
[41.07, 39.13, 39.76, 39.13, 39.16, 39.59, 39.67, 39.13, 39.06, 38.95]
[146.96]
13.517738103866577
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 20402, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.640971422195435, 'TIME_S_1KI': 0.5215651123515065, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1986.5667917442322, 'W': 146.96}
[41.07, 39.13, 39.76, 39.13, 39.16, 39.59, 39.67, 39.13, 39.06, 38.95, 39.85, 39.43, 39.2, 39.39, 39.57, 39.13, 39.11, 38.95, 44.59, 39.27]
713.5699999999999
35.6785
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 20402, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.640971422195435, 'TIME_S_1KI': 0.5215651123515065, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1986.5667917442322, 'W': 146.96, 'J_1KI': 97.3711788914926, 'W_1KI': 7.203215371042055, 'W_D': 111.28150000000001, 'J_D': 1504.2741728054286, 'W_D_1KI': 5.454440741103813, 'J_D_1KI': 0.2673483355114113}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 25477, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.081330299377441, "TIME_S_1KI": 0.39570319501422624, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1725.3062429666518, "W": 138.54, "J_1KI": 67.72014927058333, "W_1KI": 5.4378459002237305, "W_D": 102.46124999999999, "J_D": 1275.9999587640166, "W_D_1KI": 4.021715665109706, "J_D_1KI": 0.1578567203795465}

71
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 0.41213083267211914}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.5658, 0.2599, 0.8647, ..., 0.0110, 0.8951, 0.0945])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 0.41213083267211914 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '25477', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.081330299377441}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.2219, 0.1747, 0.4109, ..., 0.8392, 0.1445, 0.6192])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.081330299377441 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.2219, 0.1747, 0.4109, ..., 0.8392, 0.1445, 0.6192])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.081330299377441 seconds
[40.52, 39.06, 39.46, 40.78, 43.93, 39.17, 39.42, 39.04, 39.35, 39.02]
[138.54]
12.453488111495972
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 25477, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.081330299377441, 'TIME_S_1KI': 0.39570319501422624, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1725.3062429666518, 'W': 138.54}
[40.52, 39.06, 39.46, 40.78, 43.93, 39.17, 39.42, 39.04, 39.35, 39.02, 39.79, 46.26, 39.61, 38.94, 39.43, 40.38, 39.06, 38.95, 39.02, 40.1]
721.575
36.07875
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 25477, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.081330299377441, 'TIME_S_1KI': 0.39570319501422624, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1725.3062429666518, 'W': 138.54, 'J_1KI': 67.72014927058333, 'W_1KI': 5.4378459002237305, 'W_D': 102.46124999999999, 'J_D': 1275.9999587640166, 'W_D_1KI': 4.021715665109706, 'J_D_1KI': 0.1578567203795465}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 30984, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.317180871963501, "TIME_S_1KI": 0.33298414897894074, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1916.101525504589, "W": 149.37, "J_1KI": 61.84164489751449, "W_1KI": 4.820875290472502, "W_D": 113.45200000000001, "J_D": 1455.3494695825577, "W_D_1KI": 3.661631809966435, "J_D_1KI": 0.11817815033457382}

97
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,97 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 0.39348506927490234}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.1720, 0.8662, 0.8556, ..., 0.0402, 0.8663, 0.3929])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 0.39348506927490234 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '26684', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 9.042525291442871}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.7375, 0.5933, 0.9050, ..., 0.8578, 0.6740, 0.2052])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 9.042525291442871 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '30984', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.317180871963501}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.0633, 0.8834, 0.2857, ..., 0.1984, 0.6858, 0.2922])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.317180871963501 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.0633, 0.8834, 0.2857, ..., 0.1984, 0.6858, 0.2922])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.317180871963501 seconds
[40.3, 39.14, 39.07, 38.97, 39.41, 40.63, 44.46, 38.96, 39.02, 39.22]
[149.37]
12.827887296676636
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 30984, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.317180871963501, 'TIME_S_1KI': 0.33298414897894074, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1916.101525504589, 'W': 149.37}
[40.3, 39.14, 39.07, 38.97, 39.41, 40.63, 44.46, 38.96, 39.02, 39.22, 40.56, 39.23, 39.36, 44.65, 39.73, 38.95, 39.17, 38.94, 39.11, 39.04]
718.3599999999999
35.91799999999999
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 30984, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.317180871963501, 'TIME_S_1KI': 0.33298414897894074, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1916.101525504589, 'W': 149.37, 'J_1KI': 61.84164489751449, 'W_1KI': 4.820875290472502, 'W_D': 113.45200000000001, 'J_D': 1455.3494695825577, 'W_D_1KI': 3.661631809966435, 'J_D_1KI': 0.11817815033457382}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 30366, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.31269907951355, "TIME_S_1KI": 0.33961335307625473, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1775.0306821584702, "W": 137.59, "J_1KI": 58.454543968862225, "W_1KI": 4.531054468813805, "W_D": 102.3395, "J_D": 1320.2685696399212, "W_D_1KI": 3.370200223934664, "J_D_1KI": 0.11098597852646591}

93
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,93 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/language.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 0.3785414695739746}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.2846, 0.0684, 0.3415, ..., 0.3157, 0.0663, 0.9624])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 0.3785414695739746 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '27738', '-m', 'matrices/389000+_cols/language.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 9.591154098510742}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.1194, 0.5235, 0.5697, ..., 0.7322, 0.5132, 0.4627])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 9.591154098510742 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '30366', '-m', 'matrices/389000+_cols/language.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.31269907951355}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.4535, 0.4285, 0.5680, ..., 0.4151, 0.8586, 0.5793])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.31269907951355 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.4535, 0.4285, 0.5680, ..., 0.4151, 0.8586, 0.5793])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.31269907951355 seconds
[40.67, 39.06, 39.61, 38.65, 39.12, 40.0, 38.74, 39.55, 38.64, 38.57]
[137.59]
12.900869846343994
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 30366, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.31269907951355, 'TIME_S_1KI': 0.33961335307625473, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1775.0306821584702, 'W': 137.59}
[40.67, 39.06, 39.61, 38.65, 39.12, 40.0, 38.74, 39.55, 38.64, 38.57, 39.44, 38.8, 39.85, 39.35, 38.9, 38.8, 38.74, 39.19, 39.29, 38.76]
705.01
35.2505
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 30366, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.31269907951355, 'TIME_S_1KI': 0.33961335307625473, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1775.0306821584702, 'W': 137.59, 'J_1KI': 58.454543968862225, 'W_1KI': 4.531054468813805, 'W_D': 102.3395, 'J_D': 1320.2685696399212, 'W_D_1KI': 3.370200223934664, 'J_D_1KI': 0.11098597852646591}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 19495, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.71416425704956, "TIME_S_1KI": 0.5495852401666869, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 2110.003728497028, "W": 158.07, "J_1KI": 108.23307147971418, "W_1KI": 8.108232880225698, "W_D": 122.3625, "J_D": 1633.3607340306044, "W_D_1KI": 6.276609387022313, "J_D_1KI": 0.32195995829814383}

97
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,97 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 0.6046795845031738}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.7423, 0.4233, 0.1707, ..., 0.4030, 0.8937, 0.1151])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 0.6046795845031738 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '17364', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 9.352032661437988}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.3036, 0.7996, 0.4739, ..., 0.0238, 0.6033, 0.9918])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 9.352032661437988 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '19495', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.71416425704956}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.1117, 0.6424, 0.8924, ..., 0.5333, 0.0312, 0.4242])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.71416425704956 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.1117, 0.6424, 0.8924, ..., 0.5333, 0.0312, 0.4242])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.71416425704956 seconds
[41.06, 39.65, 39.88, 39.81, 39.59, 39.74, 39.82, 39.26, 40.0, 39.37]
[158.07]
13.34854006767273
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 19495, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.71416425704956, 'TIME_S_1KI': 0.5495852401666869, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2110.003728497028, 'W': 158.07}
[41.06, 39.65, 39.88, 39.81, 39.59, 39.74, 39.82, 39.26, 40.0, 39.37, 41.26, 39.31, 39.83, 39.61, 39.66, 39.59, 39.58, 39.11, 39.31, 39.11]
714.15
35.707499999999996
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 19495, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.71416425704956, 'TIME_S_1KI': 0.5495852401666869, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2110.003728497028, 'W': 158.07, 'J_1KI': 108.23307147971418, 'W_1KI': 8.108232880225698, 'W_D': 122.3625, 'J_D': 1633.3607340306044, 'W_D_1KI': 6.276609387022313, 'J_D_1KI': 0.32195995829814383}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 23986, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.204793214797974, "TIME_S_1KI": 0.4254478952221285, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1608.246218369007, "W": 138.89, "J_1KI": 67.04937123192725, "W_1KI": 5.79046110230968, "W_D": 103.45274999999998, "J_D": 1197.9083732981082, "W_D_1KI": 4.313047194196614, "J_D_1KI": 0.1798151919534985}

71
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 0.43773889541625977}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9846, 0.2787, 0.2893, ..., 0.3452, 0.1271, 0.5089])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 0.43773889541625977 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '23986', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.204793214797974}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.2099, 0.5726, 0.9552, ..., 0.7541, 0.8652, 0.1203])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.204793214797974 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.2099, 0.5726, 0.9552, ..., 0.7541, 0.8652, 0.1203])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.204793214797974 seconds
[40.06, 38.94, 39.06, 39.11, 39.73, 38.92, 39.35, 38.88, 39.43, 39.21]
[138.89]
11.579280138015747
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 23986, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.204793214797974, 'TIME_S_1KI': 0.4254478952221285, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1608.246218369007, 'W': 138.89}
[40.06, 38.94, 39.06, 39.11, 39.73, 38.92, 39.35, 38.88, 39.43, 39.21, 41.54, 39.4, 39.48, 39.09, 39.42, 38.94, 39.43, 39.44, 39.0, 41.44]
708.7450000000001
35.437250000000006
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 23986, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.204793214797974, 'TIME_S_1KI': 0.4254478952221285, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1608.246218369007, 'W': 138.89, 'J_1KI': 67.04937123192725, 'W_1KI': 5.79046110230968, 'W_D': 103.45274999999998, 'J_D': 1197.9083732981082, 'W_D_1KI': 4.313047194196614, 'J_D_1KI': 0.1798151919534985}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 2652, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 10.818459033966064, "TIME_S_1KI": 4.079358610092784, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1727.0440453481674, "W": 123.26, "J_1KI": 651.2232448522501, "W_1KI": 46.47812971342383, "W_D": 87.892, "J_D": 1231.4891711320877, "W_D_1KI": 33.14177978883861, "J_D_1KI": 12.49690037286524}

74
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,74 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 3.9585680961608887}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.7261, 0.8238, 0.5826, ..., 0.6988, 0.4899, 0.5621])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 3.9585680961608887 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2652', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 10.818459033966064}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.8039, 0.1348, 0.9933, ..., 0.2390, 0.5536, 0.8375])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 10.818459033966064 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.8039, 0.1348, 0.9933, ..., 0.2390, 0.5536, 0.8375])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 10.818459033966064 seconds
[40.14, 39.06, 39.78, 39.45, 39.62, 39.78, 39.17, 39.08, 39.09, 38.93]
[123.26]
14.011390924453735
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2652, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 10.818459033966064, 'TIME_S_1KI': 4.079358610092784, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1727.0440453481674, 'W': 123.26}
[40.14, 39.06, 39.78, 39.45, 39.62, 39.78, 39.17, 39.08, 39.09, 38.93, 39.84, 39.42, 39.65, 39.42, 38.97, 39.03, 39.01, 38.92, 39.07, 38.77]
707.36
35.368
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2652, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 10.818459033966064, 'TIME_S_1KI': 4.079358610092784, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1727.0440453481674, 'W': 123.26, 'J_1KI': 651.2232448522501, 'W_1KI': 46.47812971342383, 'W_D': 87.892, 'J_D': 1231.4891711320877, 'W_D_1KI': 33.14177978883861, 'J_D_1KI': 12.49690037286524}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 8118, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.907745361328125, "TIME_S_1KI": 1.3436493423661153, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1301.0154501628876, "W": 88.97, "J_1KI": 160.2630512642138, "W_1KI": 10.959595959595958, "W_D": 72.9665, "J_D": 1066.9949853243827, "W_D_1KI": 8.988236018723823, "J_D_1KI": 1.1071983270169774}

71
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,71 @@
['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', '1000', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 1.293351411819458}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.2979, 0.5597, 0.8769, ..., 0.0942, 0.8424, 0.4292])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 1.293351411819458 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', '8118', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.907745361328125}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6448, 0.4532, 0.0841, ..., 0.5791, 0.6160, 0.9399])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.907745361328125 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6448, 0.4532, 0.0841, ..., 0.5791, 0.6160, 0.9399])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.907745361328125 seconds
[18.52, 17.54, 17.79, 17.99, 17.73, 17.74, 17.74, 17.7, 17.79, 17.98]
[88.97]
14.623080253601074
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 8118, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.907745361328125, 'TIME_S_1KI': 1.3436493423661153, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1301.0154501628876, 'W': 88.97}
[18.52, 17.54, 17.79, 17.99, 17.73, 17.74, 17.74, 17.7, 17.79, 17.98, 18.52, 17.6, 17.66, 17.43, 18.0, 17.58, 17.75, 17.82, 17.89, 17.62]
320.06999999999994
16.003499999999995
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 8118, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.907745361328125, 'TIME_S_1KI': 1.3436493423661153, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1301.0154501628876, 'W': 88.97, 'J_1KI': 160.2630512642138, 'W_1KI': 10.959595959595958, 'W_D': 72.9665, 'J_D': 1066.9949853243827, 'W_D_1KI': 8.988236018723823, 'J_D_1KI': 1.1071983270169774}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 13615, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.780885696411133, "TIME_S_1KI": 0.7918388319068037, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1298.2887223243713, "W": 89.8, "J_1KI": 95.35723263491526, "W_1KI": 6.595666544252662, "W_D": 73.5375, "J_D": 1063.172682827711, "W_D_1KI": 5.401211898641204, "J_D_1KI": 0.3967103855043117}

71
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,71 @@
['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', '1000', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 0.7711856365203857}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.3467, 0.9628, 0.5083, ..., 0.1832, 0.8742, 0.8835])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 0.7711856365203857 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', '13615', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.780885696411133}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0947, 0.1039, 0.1767, ..., 0.1078, 0.6970, 0.7249])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.780885696411133 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0947, 0.1039, 0.1767, ..., 0.1078, 0.6970, 0.7249])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.780885696411133 seconds
[19.22, 17.93, 18.06, 17.51, 17.76, 17.39, 17.9, 17.91, 17.46, 17.78]
[89.8]
14.457558155059814
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13615, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.780885696411133, 'TIME_S_1KI': 0.7918388319068037, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1298.2887223243713, 'W': 89.8}
[19.22, 17.93, 18.06, 17.51, 17.76, 17.39, 17.9, 17.91, 17.46, 17.78, 18.08, 17.65, 17.61, 17.86, 17.77, 17.72, 21.98, 18.52, 17.82, 17.72]
325.25
16.2625
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13615, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.780885696411133, 'TIME_S_1KI': 0.7918388319068037, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1298.2887223243713, 'W': 89.8, 'J_1KI': 95.35723263491526, 'W_1KI': 6.595666544252662, 'W_D': 73.5375, 'J_D': 1063.172682827711, 'W_D_1KI': 5.401211898641204, 'J_D_1KI': 0.3967103855043117}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 12165, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.51817512512207, "TIME_S_1KI": 0.8646259864465327, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1289.3463970065116, "W": 89.63, "J_1KI": 105.98819539716494, "W_1KI": 7.3678586107685975, "W_D": 73.37299999999999, "J_D": 1055.4860335552692, "W_D_1KI": 6.031483764899301, "J_D_1KI": 0.49580631030820393}

74
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,74 @@
['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', '1000', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 0.8631081581115723}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.5077, 0.7328, 0.3933, ..., 0.4074, 0.1030, 0.0500])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 0.8631081581115723 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', '12165', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.51817512512207}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9100, 0.1506, 0.3829, ..., 0.6719, 0.7400, 0.8631])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.51817512512207 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9100, 0.1506, 0.3829, ..., 0.6719, 0.7400, 0.8631])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.51817512512207 seconds
[18.23, 17.69, 18.03, 22.07, 17.84, 17.63, 17.63, 17.56, 17.59, 17.97]
[89.63]
14.385210275650024
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 12165, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.51817512512207, 'TIME_S_1KI': 0.8646259864465327, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1289.3463970065116, 'W': 89.63}
[18.23, 17.69, 18.03, 22.07, 17.84, 17.63, 17.63, 17.56, 17.59, 17.97, 18.65, 17.64, 17.75, 18.23, 17.75, 18.05, 17.76, 17.53, 18.1, 17.73]
325.14
16.256999999999998
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 12165, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.51817512512207, 'TIME_S_1KI': 0.8646259864465327, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1289.3463970065116, 'W': 89.63, 'J_1KI': 105.98819539716494, 'W_1KI': 7.3678586107685975, 'W_D': 73.37299999999999, 'J_D': 1055.4860335552692, 'W_D_1KI': 6.031483764899301, 'J_D_1KI': 0.49580631030820393}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 13328, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.746992826461792, "TIME_S_1KI": 0.8063470007849484, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1285.765242910385, "W": 89.74, "J_1KI": 96.47098161092326, "W_1KI": 6.733193277310924, "W_D": 73.49574999999999, "J_D": 1053.0229646939038, "W_D_1KI": 5.51438700480192, "J_D_1KI": 0.4137445231694118}

71
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,71 @@
['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', '1000', '-m', 'matrices/389000+_cols/language.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 0.7877652645111084}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.0377, 0.4846, 0.1087, ..., 0.8181, 0.0416, 0.1571])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 0.7877652645111084 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', '13328', '-m', 'matrices/389000+_cols/language.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.746992826461792}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.5687, 0.0271, 0.0300, ..., 0.3524, 0.7739, 0.4785])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.746992826461792 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.5687, 0.0271, 0.0300, ..., 0.3524, 0.7739, 0.4785])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.746992826461792 seconds
[18.42, 17.72, 17.91, 18.41, 17.89, 17.65, 19.38, 18.4, 18.2, 17.87]
[89.74]
14.327671527862549
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13328, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.746992826461792, 'TIME_S_1KI': 0.8063470007849484, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1285.765242910385, 'W': 89.74}
[18.42, 17.72, 17.91, 18.41, 17.89, 17.65, 19.38, 18.4, 18.2, 17.87, 19.34, 17.9, 18.16, 17.6, 17.86, 17.8, 17.9, 17.65, 17.71, 17.86]
324.885
16.24425
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13328, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.746992826461792, 'TIME_S_1KI': 0.8063470007849484, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1285.765242910385, 'W': 89.74, 'J_1KI': 96.47098161092326, 'W_1KI': 6.733193277310924, 'W_D': 73.49574999999999, 'J_D': 1053.0229646939038, 'W_D_1KI': 5.51438700480192, 'J_D_1KI': 0.4137445231694118}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 5897, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.55986738204956, "TIME_S_1KI": 1.790718565719783, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1299.1635441350936, "W": 89.05999999999999, "J_1KI": 220.30923251400603, "W_1KI": 15.102594539596403, "W_D": 72.89299999999999, "J_D": 1063.327287476301, "W_D_1KI": 12.361031032728503, "J_D_1KI": 2.096155847503562}

74
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,74 @@
['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', '1000', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 1.7802655696868896}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.5518, 0.1807, 0.4021, ..., 0.3397, 0.2107, 0.4589])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 1.7802655696868896 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', '5897', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.55986738204956}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.5366, 0.8162, 0.5634, ..., 0.9410, 0.0469, 0.5938])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.55986738204956 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.5366, 0.8162, 0.5634, ..., 0.9410, 0.0469, 0.5938])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.55986738204956 seconds
[18.26, 17.65, 17.89, 17.69, 17.78, 17.53, 17.59, 17.98, 17.75, 17.71]
[89.06]
14.587508916854858
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 5897, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.55986738204956, 'TIME_S_1KI': 1.790718565719783, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1299.1635441350936, 'W': 89.05999999999999}
[18.26, 17.65, 17.89, 17.69, 17.78, 17.53, 17.59, 17.98, 17.75, 17.71, 18.09, 17.83, 17.84, 17.57, 17.77, 17.94, 21.22, 17.67, 17.65, 17.92]
323.34000000000003
16.167
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 5897, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.55986738204956, 'TIME_S_1KI': 1.790718565719783, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1299.1635441350936, 'W': 89.05999999999999, 'J_1KI': 220.30923251400603, 'W_1KI': 15.102594539596403, 'W_D': 72.89299999999999, 'J_D': 1063.327287476301, 'W_D_1KI': 12.361031032728503, 'J_D_1KI': 2.096155847503562}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 13697, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.777354717254639, "TIME_S_1KI": 0.7868405283824662, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1316.45538356781, "W": 89.82, "J_1KI": 96.11268040941886, "W_1KI": 6.557640359202745, "W_D": 73.69574999999999, "J_D": 1080.128777928829, "W_D_1KI": 5.3804300211725185, "J_D_1KI": 0.3928181369038854}

71
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,71 @@
['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', '1000', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 0.7665784358978271}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.5208, 0.6283, 0.9927, ..., 0.7747, 0.7207, 0.3302])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 0.7665784358978271 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', '13697', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.777354717254639}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9293, 0.9172, 0.4372, ..., 0.0528, 0.4444, 0.3291])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.777354717254639 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9293, 0.9172, 0.4372, ..., 0.0528, 0.4444, 0.3291])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.777354717254639 seconds
[18.67, 17.76, 17.77, 17.98, 17.83, 17.63, 17.99, 17.72, 18.01, 17.56]
[89.82]
14.656595230102539
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13697, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.777354717254639, 'TIME_S_1KI': 0.7868405283824662, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1316.45538356781, 'W': 89.82}
[18.67, 17.76, 17.77, 17.98, 17.83, 17.63, 17.99, 17.72, 18.01, 17.56, 18.11, 17.79, 18.09, 17.72, 17.66, 18.49, 17.86, 18.34, 17.71, 17.93]
322.485
16.12425
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 13697, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.777354717254639, 'TIME_S_1KI': 0.7868405283824662, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1316.45538356781, 'W': 89.82, 'J_1KI': 96.11268040941886, 'W_1KI': 6.557640359202745, 'W_D': 73.69574999999999, 'J_D': 1080.128777928829, 'W_D_1KI': 5.3804300211725185, 'J_D_1KI': 0.3928181369038854}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 1887, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 10.573626518249512, "TIME_S_1KI": 5.603405680047436, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1443.276729001999, "W": 85.01, "J_1KI": 764.8525325924743, "W_1KI": 45.050344462109166, "W_D": 68.79925, "J_D": 1168.055011149168, "W_D_1KI": 36.459591944886064, "J_D_1KI": 19.321458370368873}

74
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,74 @@
['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', '1000', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 5.56341028213501}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.2549, 0.6086, 0.7138, ..., 0.3139, 0.6424, 0.7605])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 5.56341028213501 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', '1887', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 10.573626518249512}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3061, 0.1365, 0.6683, ..., 0.9071, 0.4159, 0.8227])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 10.573626518249512 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3061, 0.1365, 0.6683, ..., 0.9071, 0.4159, 0.8227])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 10.573626518249512 seconds
[18.08, 17.49, 17.66, 17.63, 17.95, 17.53, 17.51, 17.68, 17.93, 17.61]
[85.01]
16.977728843688965
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1887, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 10.573626518249512, 'TIME_S_1KI': 5.603405680047436, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1443.276729001999, 'W': 85.01}
[18.08, 17.49, 17.66, 17.63, 17.95, 17.53, 17.51, 17.68, 17.93, 17.61, 18.14, 17.76, 17.8, 22.06, 18.45, 17.61, 17.74, 18.01, 17.71, 17.56]
324.21500000000003
16.21075
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1887, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 10.573626518249512, 'TIME_S_1KI': 5.603405680047436, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1443.276729001999, 'W': 85.01, 'J_1KI': 764.8525325924743, 'W_1KI': 45.050344462109166, 'W_D': 68.79925, 'J_D': 1168.055011149168, 'W_D_1KI': 36.459591944886064, 'J_D_1KI': 19.321458370368873}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 88.16403460502625, "TIME_S_1KI": 88.16403460502625, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 2373.6426183891294, "W": 25.696559171496023, "J_1KI": 2373.6426183891294, "W_1KI": 25.696559171496023, "W_D": 4.897559171496024, "J_D": 452.3973461956977, "W_D_1KI": 4.897559171496024, "J_D_1KI": 4.897559171496024}

49
pytorch/output_389000+_1core/altra_1_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/amazon0312.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 88.16403460502625}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.0536, 0.3063, 0.9163, ..., 0.7704, 0.9774, 0.8233])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 88.16403460502625 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.0536, 0.3063, 0.9163, ..., 0.7704, 0.9774, 0.8233])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 88.16403460502625 seconds
[22.96, 23.04, 23.08, 23.12, 23.16, 23.24, 23.24, 22.92, 22.8, 22.84]
[22.88, 23.08, 23.44, 27.52, 29.2, 30.24, 30.68, 28.56, 27.76, 26.76, 26.68, 26.84, 26.68, 26.6, 26.56, 26.56, 26.52, 26.72, 26.88, 26.8, 26.92, 26.88, 26.8, 27.0, 27.0, 26.96, 27.08, 26.92, 27.04, 27.08, 27.0, 26.88, 26.72, 26.84, 26.96, 27.04, 27.16, 27.16, 27.28, 27.44, 27.32, 27.44, 27.6, 27.56, 27.44, 27.44, 27.36, 27.72, 27.84, 28.0, 27.96, 27.92, 27.6, 27.56, 27.68, 27.4, 27.4, 27.12, 27.12, 27.12, 27.28, 27.32, 27.36, 27.24, 27.2, 27.16, 27.24, 27.2, 27.24, 27.32, 27.24, 27.28, 27.2, 26.92, 26.92, 27.2, 27.24, 27.32, 27.24, 27.24, 26.96, 26.96, 27.04, 26.84, 26.88, 26.96, 27.4, 27.48]
92.37200212478638
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 88.16403460502625, 'TIME_S_1KI': 88.16403460502625, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2373.6426183891294, 'W': 25.696559171496023}
[22.96, 23.04, 23.08, 23.12, 23.16, 23.24, 23.24, 22.92, 22.8, 22.84, 23.16, 22.88, 22.76, 23.04, 23.12, 23.44, 23.56, 23.36, 23.24, 23.0]
415.98
20.799
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 88.16403460502625, 'TIME_S_1KI': 88.16403460502625, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2373.6426183891294, 'W': 25.696559171496023, 'J_1KI': 2373.6426183891294, 'W_1KI': 25.696559171496023, 'W_D': 4.897559171496024, 'J_D': 452.3973461956977, 'W_D_1KI': 4.897559171496024, 'J_D_1KI': 4.897559171496024}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 53.30746507644653, "TIME_S_1KI": 53.30746507644653, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1536.7467108154294, "W": 25.681407272182962, "J_1KI": 1536.7467108154294, "W_1KI": 25.681407272182962, "W_D": 4.658407272182959, "J_D": 278.7538851470942, "W_D_1KI": 4.658407272182959, "J_D_1KI": 4.658407272182959}

49
pytorch/output_389000+_1core/altra_1_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/darcy003.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 53.30746507644653}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.6125, 0.5278, 0.3887, ..., 0.3141, 0.8902, 0.1690])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 53.30746507644653 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.6125, 0.5278, 0.3887, ..., 0.3141, 0.8902, 0.1690])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 53.30746507644653 seconds
[23.16, 23.4, 23.4, 23.32, 23.48, 23.4, 23.36, 23.56, 23.24, 23.28]
[23.32, 22.84, 26.36, 27.44, 28.48, 28.48, 29.44, 30.12, 27.2, 26.6, 26.92, 27.16, 27.24, 27.2, 27.28, 27.12, 27.12, 26.96, 27.08, 27.04, 26.48, 26.72, 26.84, 26.76, 26.96, 27.12, 26.68, 26.68, 26.84, 26.64, 26.8, 26.96, 26.72, 26.6, 26.64, 26.44, 26.36, 26.68, 26.8, 26.84, 27.16, 27.28, 27.36, 27.0, 27.0, 27.12, 27.04, 26.96, 26.88, 26.88, 26.88, 26.64, 27.32, 27.88, 28.04, 28.0, 27.64]
59.83888244628906
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 53.30746507644653, 'TIME_S_1KI': 53.30746507644653, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1536.7467108154294, 'W': 25.681407272182962}
[23.16, 23.4, 23.4, 23.32, 23.48, 23.4, 23.36, 23.56, 23.24, 23.28, 23.36, 23.32, 23.28, 23.32, 23.28, 23.36, 23.4, 23.48, 23.32, 23.28]
420.46000000000004
21.023000000000003
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 53.30746507644653, 'TIME_S_1KI': 53.30746507644653, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1536.7467108154294, 'W': 25.681407272182962, 'J_1KI': 1536.7467108154294, 'W_1KI': 25.681407272182962, 'W_D': 4.658407272182959, 'J_D': 278.7538851470942, 'W_D_1KI': 4.658407272182959, 'J_D_1KI': 4.658407272182959}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 62.08789658546448, "TIME_S_1KI": 62.08789658546448, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1791.450892791748, "W": 25.109871656394702, "J_1KI": 1791.450892791748, "W_1KI": 25.109871656394702, "W_D": 4.4798716563947, "J_D": 319.61414173126184, "W_D_1KI": 4.4798716563947, "J_D_1KI": 4.4798716563947}

51
pytorch/output_389000+_1core/altra_1_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/helm2d03.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 62.08789658546448}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.2732, 0.1117, 0.4132, ..., 0.8859, 0.7833, 0.1406])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 62.08789658546448 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.2732, 0.1117, 0.4132, ..., 0.8859, 0.7833, 0.1406])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 62.08789658546448 seconds
[23.16, 22.92, 22.92, 22.96, 22.84, 23.04, 23.12, 23.2, 23.36, 23.44]
[23.36, 23.32, 23.32, 24.2, 25.04, 27.48, 28.36, 28.84, 28.4, 27.24, 26.92, 26.76, 26.72, 26.8, 26.68, 26.6, 26.6, 26.56, 26.52, 26.56, 26.48, 26.48, 26.44, 26.6, 26.6, 26.56, 26.4, 26.48, 26.56, 26.36, 26.28, 26.2, 26.44, 26.6, 26.68, 27.2, 27.2, 27.04, 27.04, 26.84, 26.64, 26.64, 26.44, 26.4, 26.48, 26.48, 26.16, 26.52, 26.72, 26.8, 26.84, 26.84, 26.72, 26.64, 26.6, 26.8, 26.72, 26.8, 26.96, 27.0, 27.2, 27.32, 27.16, 26.88, 26.72, 26.68, 26.68, 26.68]
71.34448623657227
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 62.08789658546448, 'TIME_S_1KI': 62.08789658546448, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1791.450892791748, 'W': 25.109871656394702}
[23.16, 22.92, 22.92, 22.96, 22.84, 23.04, 23.12, 23.2, 23.36, 23.44, 23.08, 23.08, 23.04, 22.76, 22.48, 22.48, 22.48, 22.64, 22.96, 22.96]
412.6
20.630000000000003
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 62.08789658546448, 'TIME_S_1KI': 62.08789658546448, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1791.450892791748, 'W': 25.109871656394702, 'J_1KI': 1791.450892791748, 'W_1KI': 25.109871656394702, 'W_D': 4.4798716563947, 'J_D': 319.61414173126184, 'W_D_1KI': 4.4798716563947, 'J_D_1KI': 4.4798716563947}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 32.617069482803345, "TIME_S_1KI": 32.617069482803345, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 877.9408810043335, "W": 25.358346526187013, "J_1KI": 877.9408810043335, "W_1KI": 25.358346526187013, "W_D": 4.655346526187017, "J_D": 161.17450821805022, "W_D_1KI": 4.655346526187017, "J_D_1KI": 4.655346526187017}

49
pytorch/output_389000+_1core/altra_1_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/language.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 32.617069482803345}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.1808, 0.0389, 0.7706, ..., 0.1715, 0.5157, 0.4224])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 32.617069482803345 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.1808, 0.0389, 0.7706, ..., 0.1715, 0.5157, 0.4224])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 32.617069482803345 seconds
[23.16, 23.16, 22.96, 23.16, 22.92, 22.92, 23.0, 22.92, 22.84, 23.16]
[23.24, 23.4, 24.8, 25.6, 27.28, 28.0, 28.0, 28.48, 27.76, 27.96, 26.76, 26.64, 26.76, 26.88, 26.88, 26.96, 26.96, 27.2, 27.0, 27.12, 27.08, 26.68, 27.08, 27.16, 27.08, 27.08, 27.28, 27.0, 27.0, 27.24, 27.16, 27.36, 27.36]
34.62137722969055
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 32.617069482803345, 'TIME_S_1KI': 32.617069482803345, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 877.9408810043335, 'W': 25.358346526187013}
[23.16, 23.16, 22.96, 23.16, 22.92, 22.92, 23.0, 22.92, 22.84, 23.16, 22.96, 23.28, 23.08, 23.16, 23.04, 22.88, 22.76, 22.88, 22.96, 23.0]
414.05999999999995
20.702999999999996
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 32.617069482803345, 'TIME_S_1KI': 32.617069482803345, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 877.9408810043335, 'W': 25.358346526187013, 'J_1KI': 877.9408810043335, 'W_1KI': 25.358346526187013, 'W_D': 4.655346526187017, 'J_D': 161.17450821805022, 'W_D_1KI': 4.655346526187017, 'J_D_1KI': 4.655346526187017}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 136.75263905525208, "TIME_S_1KI": 136.75263905525208, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 3809.28440010071, "W": 25.30736544046405, "J_1KI": 3809.28440010071, "W_1KI": 25.30736544046405, "W_D": 4.417365440464049, "J_D": 664.905294132235, "W_D_1KI": 4.417365440464049, "J_D_1KI": 4.417365440464049}

51
pytorch/output_389000+_1core/altra_1_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/marine1.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 136.75263905525208}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.1033, 0.2543, 0.2854, ..., 0.8643, 0.3799, 0.0773])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 136.75263905525208 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.1033, 0.2543, 0.2854, ..., 0.8643, 0.3799, 0.0773])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 136.75263905525208 seconds
[23.6, 23.2, 23.36, 23.44, 23.4, 23.52, 23.52, 23.4, 23.16, 22.92]
[23.12, 23.0, 23.0, 25.76, 27.64, 28.96, 29.28, 28.32, 27.36, 27.36, 27.0, 26.64, 26.6, 26.6, 26.76, 26.64, 26.6, 26.6, 26.4, 26.48, 26.56, 26.76, 26.96, 27.0, 27.04, 26.84, 26.92, 26.64, 26.64, 26.8, 26.88, 26.88, 26.88, 27.04, 27.12, 26.96, 26.96, 26.84, 26.6, 26.6, 26.24, 26.04, 26.0, 26.24, 26.48, 26.88, 27.2, 27.36, 27.16, 27.24, 27.16, 27.04, 27.04, 27.04, 26.88, 26.8, 26.48, 26.68, 26.64, 26.72, 26.68, 26.84, 26.6, 26.48, 26.36, 26.4, 26.28, 26.48, 26.6, 26.48, 26.48, 26.8, 26.76, 26.72, 26.76, 26.76, 26.68, 26.52, 26.48, 26.56, 26.6, 26.52, 26.28, 26.32, 26.24, 26.44, 26.52, 26.6, 26.44, 26.52, 26.56, 26.4, 26.56, 26.64, 26.92, 27.04, 26.92, 26.92, 26.96, 26.92, 26.84, 26.88, 27.16, 27.08, 26.96, 26.72, 26.56, 26.44, 26.4, 26.4, 26.64, 26.72, 26.84, 26.96, 27.2, 26.88, 26.8, 26.8, 26.92, 26.92, 27.08, 27.28, 27.2, 27.0, 26.92, 26.8, 26.64, 26.92, 26.72, 26.76, 26.8, 26.8, 26.76, 26.84, 26.88, 26.72, 26.6, 26.64, 26.52, 26.8, 26.96, 26.96, 27.16]
150.5207805633545
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 136.75263905525208, 'TIME_S_1KI': 136.75263905525208, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 3809.28440010071, 'W': 25.30736544046405}
[23.6, 23.2, 23.36, 23.44, 23.4, 23.52, 23.52, 23.4, 23.16, 22.92, 23.28, 23.28, 23.28, 23.16, 23.08, 22.96, 22.84, 22.8, 22.92, 23.16]
417.8
20.89
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 136.75263905525208, 'TIME_S_1KI': 136.75263905525208, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 3809.28440010071, 'W': 25.30736544046405, 'J_1KI': 3809.28440010071, 'W_1KI': 25.30736544046405, 'W_D': 4.417365440464049, 'J_D': 664.905294132235, 'W_D_1KI': 4.417365440464049, 'J_D_1KI': 4.417365440464049}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 51.17483377456665, "TIME_S_1KI": 51.17483377456665, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1410.4799920654298, "W": 25.376735316064863, "J_1KI": 1410.4799920654298, "W_1KI": 25.376735316064863, "W_D": 4.522735316064864, "J_D": 251.3809437370302, "W_D_1KI": 4.522735316064864, "J_D_1KI": 4.522735316064864}

49
pytorch/output_389000+_1core/altra_1_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/mario002.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 51.17483377456665}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0328, 0.3187, 0.7172, ..., 0.3931, 0.0888, 0.1198])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 51.17483377456665 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0328, 0.3187, 0.7172, ..., 0.3931, 0.0888, 0.1198])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 51.17483377456665 seconds
[23.16, 23.16, 23.16, 23.28, 23.28, 23.48, 23.44, 23.4, 23.44, 23.44]
[23.4, 23.48, 23.92, 24.84, 26.8, 27.6, 28.64, 28.08, 28.08, 26.88, 26.8, 26.8, 26.76, 26.76, 26.8, 26.64, 26.68, 26.48, 26.72, 26.96, 27.12, 27.2, 27.08, 27.12, 27.2, 27.12, 26.88, 27.04, 27.04, 27.2, 27.36, 27.24, 27.2, 27.2, 27.12, 27.12, 27.2, 26.92, 26.88, 26.88, 26.92, 26.84, 26.6, 26.84, 26.68, 26.72, 26.96, 26.72, 26.64, 26.84, 26.76, 27.0, 27.2]
55.58161735534668
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 51.17483377456665, 'TIME_S_1KI': 51.17483377456665, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1410.4799920654298, 'W': 25.376735316064863}
[23.16, 23.16, 23.16, 23.28, 23.28, 23.48, 23.44, 23.4, 23.44, 23.44, 23.08, 22.96, 22.96, 23.04, 23.08, 22.96, 23.04, 23.12, 22.88, 23.12]
417.08
20.854
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 51.17483377456665, 'TIME_S_1KI': 51.17483377456665, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1410.4799920654298, 'W': 25.376735316064863, 'J_1KI': 1410.4799920654298, 'W_1KI': 25.376735316064863, 'W_D': 4.522735316064864, 'J_D': 251.3809437370302, 'W_D_1KI': 4.522735316064864, 'J_D_1KI': 4.522735316064864}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 283.81978273391724, "TIME_S_1KI": 283.81978273391724, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 7375.577878112791, "W": 25.373727151984518, "J_1KI": 7375.577878112791, "W_1KI": 25.373727151984518, "W_D": 4.492727151984518, "J_D": 1305.9358129017337, "W_D_1KI": 4.492727151984518, "J_D_1KI": 4.492727151984518}

51
pytorch/output_389000+_1core/altra_1_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/test1.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 283.81978273391724}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.4069, 0.7437, 0.0499, ..., 0.6858, 0.0232, 0.7224])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 283.81978273391724 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.4069, 0.7437, 0.0499, ..., 0.6858, 0.0232, 0.7224])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 283.81978273391724 seconds
[22.72, 22.6, 22.72, 22.52, 22.8, 23.2, 23.32, 23.36, 23.36, 23.36]
[23.36, 23.0, 23.52, 26.52, 28.0, 29.08, 29.6, 28.4, 27.72, 27.56, 27.4, 26.84, 27.08, 27.28, 27.24, 27.04, 26.76, 26.76, 26.68, 26.48, 26.44, 26.48, 26.36, 26.2, 26.48, 26.48, 26.64, 26.64, 26.68, 26.52, 26.52, 26.68, 26.64, 26.68, 26.68, 26.68, 26.64, 26.64, 26.64, 26.92, 26.84, 26.6, 26.72, 26.68, 26.44, 26.76, 26.96, 26.8, 26.8, 26.96, 27.0, 26.92, 26.96, 26.96, 26.64, 26.52, 26.36, 26.56, 26.6, 26.72, 26.72, 26.72, 26.64, 26.6, 26.52, 26.64, 26.52, 26.48, 26.44, 26.44, 26.48, 26.56, 26.76, 26.64, 26.76, 26.72, 26.76, 26.8, 26.76, 26.64, 26.44, 26.44, 26.44, 26.44, 26.64, 26.56, 26.76, 26.76, 26.68, 26.68, 26.72, 26.72, 26.8, 26.8, 26.64, 26.4, 26.52, 26.56, 26.68, 26.68, 26.96, 26.72, 26.64, 26.76, 26.76, 26.84, 26.76, 26.96, 26.8, 26.76, 26.68, 26.72, 26.8, 27.12, 27.08, 26.92, 26.68, 26.28, 26.2, 26.44, 26.4, 26.68, 26.88, 26.92, 27.08, 26.96, 26.96, 26.88, 26.84, 26.68, 26.64, 26.68, 26.52, 26.48, 26.28, 26.28, 26.24, 26.44, 26.76, 27.04, 26.92, 27.0, 27.0, 27.0, 26.84, 26.92, 26.88, 26.92, 26.92, 26.8, 26.8, 26.84, 26.92, 26.92, 27.08, 26.92, 27.0, 26.92, 26.8, 26.8, 26.96, 26.84, 26.92, 26.84, 26.68, 26.6, 26.68, 26.72, 26.96, 26.76, 26.76, 26.6, 26.4, 26.28, 26.2, 26.44, 26.8, 26.76, 26.92, 27.04, 26.8, 26.6, 26.6, 26.32, 26.24, 26.4, 26.28, 26.36, 26.64, 26.64, 26.6, 26.68, 26.68, 26.44, 26.24, 26.12, 26.36, 26.44, 26.68, 26.8, 26.92, 26.96, 26.92, 26.88, 26.88, 26.76, 26.76, 26.8, 26.68, 26.72, 26.64, 26.52, 26.64, 26.92, 26.92, 26.76, 26.6, 26.6, 26.56, 26.36, 26.36, 26.64, 26.6, 26.72, 26.76, 26.84, 26.72, 27.04, 26.72, 26.6, 26.88, 26.76, 26.72, 26.64, 26.8, 27.04, 27.04, 26.96, 27.04, 27.04, 26.8, 26.68, 26.88, 26.84, 27.0, 26.92, 26.96, 27.0, 26.76, 26.8, 26.88, 26.88, 26.96, 27.0, 27.12, 27.32, 27.32, 27.32, 27.32, 27.24, 26.88, 26.6, 26.68, 26.64, 26.92, 27.08, 27.28, 27.04, 26.92, 27.08, 27.4, 27.56, 27.6, 27.48, 27.2, 26.88]
290.67774844169617
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 283.81978273391724, 'TIME_S_1KI': 283.81978273391724, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 7375.577878112791, 'W': 25.373727151984518}
[22.72, 22.6, 22.72, 22.52, 22.8, 23.2, 23.32, 23.36, 23.36, 23.36, 23.48, 23.28, 23.2, 23.24, 23.24, 23.32, 23.6, 23.8, 23.52, 23.52]
417.62
20.881
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 283.81978273391724, 'TIME_S_1KI': 283.81978273391724, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 7375.577878112791, 'W': 25.373727151984518, 'J_1KI': 7375.577878112791, 'W_1KI': 25.373727151984518, 'W_D': 4.492727151984518, 'J_D': 1305.9358129017337, 'W_D_1KI': 4.492727151984518, 'J_D_1KI': 4.492727151984518}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 1665, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.677687168121338, "TIME_S_1KI": 6.413025326199002, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 984.0334419322014, "W": 72.51, "J_1KI": 591.0110762355564, "W_1KI": 43.54954954954955, "W_D": 37.19775000000001, "J_D": 504.81078423160335, "W_D_1KI": 22.340990990991, "J_D_1KI": 13.418012607201801}

71
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 6.305053472518921}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6406, 0.1468, 0.7280, ..., 0.0181, 0.2043, 0.6040])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 6.305053472518921 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1665', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.677687168121338}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.7736, 0.6337, 0.2989, ..., 0.0625, 0.4089, 0.3233])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.677687168121338 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.7736, 0.6337, 0.2989, ..., 0.0625, 0.4089, 0.3233])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 10.677687168121338 seconds
[40.11, 38.6, 39.5, 38.36, 38.75, 38.98, 39.85, 38.73, 38.93, 38.39]
[72.51]
13.571003198623657
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1665, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.677687168121338, 'TIME_S_1KI': 6.413025326199002, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 984.0334419322014, 'W': 72.51}
[40.11, 38.6, 39.5, 38.36, 38.75, 38.98, 39.85, 38.73, 38.93, 38.39, 39.83, 38.29, 40.69, 40.97, 41.34, 39.2, 38.48, 38.73, 38.45, 38.46]
706.2449999999999
35.31224999999999
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1665, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 10.677687168121338, 'TIME_S_1KI': 6.413025326199002, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 984.0334419322014, 'W': 72.51, 'J_1KI': 591.0110762355564, 'W_1KI': 43.54954954954955, 'W_D': 37.19775000000001, 'J_D': 504.81078423160335, 'W_D_1KI': 22.340990990991, 'J_D_1KI': 13.418012607201801}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 2770, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.539327144622803, "TIME_S_1KI": 3.804811243546138, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 940.3520177507401, "W": 70.14, "J_1KI": 339.477262725899, "W_1KI": 25.32129963898917, "W_D": 35.342749999999995, "J_D": 473.8327099424004, "W_D_1KI": 12.759115523465702, "J_D_1KI": 4.606178889337799}

71
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 3.790585994720459}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.8724, 0.5946, 0.8360, ..., 0.1630, 0.5271, 0.0708])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 3.790585994720459 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2770', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.539327144622803}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.4798, 0.6348, 0.4010, ..., 0.9410, 0.2128, 0.7861])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.539327144622803 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.4798, 0.6348, 0.4010, ..., 0.9410, 0.2128, 0.7861])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.539327144622803 seconds
[39.64, 38.41, 38.51, 38.38, 38.48, 38.55, 38.6, 38.35, 39.22, 38.54]
[70.14]
13.406786680221558
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2770, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.539327144622803, 'TIME_S_1KI': 3.804811243546138, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 940.3520177507401, 'W': 70.14}
[39.64, 38.41, 38.51, 38.38, 38.48, 38.55, 38.6, 38.35, 39.22, 38.54, 39.52, 38.45, 38.63, 38.39, 38.55, 38.9, 38.88, 38.72, 38.83, 38.49]
695.945
34.797250000000005
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2770, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.539327144622803, 'TIME_S_1KI': 3.804811243546138, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 940.3520177507401, 'W': 70.14, 'J_1KI': 339.477262725899, 'W_1KI': 25.32129963898917, 'W_D': 35.342749999999995, 'J_D': 473.8327099424004, 'W_D_1KI': 12.759115523465702, 'J_D_1KI': 4.606178889337799}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 2881, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.489487886428833, "TIME_S_1KI": 3.640919085882969, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 982.9177148509026, "W": 72.89, "J_1KI": 341.1724105695601, "W_1KI": 25.30024297119056, "W_D": 38.247499999999995, "J_D": 515.765472612977, "W_D_1KI": 13.275772301284276, "J_D_1KI": 4.6080431451871835}

74
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,74 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 3.64394474029541}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.2874, 0.3706, 0.3465, ..., 0.0468, 0.1058, 0.2863])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 3.64394474029541 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2881', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.489487886428833}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.4950, 0.0177, 0.5787, ..., 0.9424, 0.3532, 0.5521])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.489487886428833 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.4950, 0.0177, 0.5787, ..., 0.9424, 0.3532, 0.5521])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.489487886428833 seconds
[39.75, 38.21, 38.34, 38.21, 38.27, 38.42, 38.28, 38.17, 38.78, 38.77]
[72.89]
13.484946012496948
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2881, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.489487886428833, 'TIME_S_1KI': 3.640919085882969, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 982.9177148509026, 'W': 72.89}
[39.75, 38.21, 38.34, 38.21, 38.27, 38.42, 38.28, 38.17, 38.78, 38.77, 39.12, 38.21, 38.45, 38.24, 39.01, 38.22, 38.93, 38.26, 38.73, 38.6]
692.8500000000001
34.642500000000005
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2881, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.489487886428833, 'TIME_S_1KI': 3.640919085882969, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 982.9177148509026, 'W': 72.89, 'J_1KI': 341.1724105695601, 'W_1KI': 25.30024297119056, 'W_D': 38.247499999999995, 'J_D': 515.765472612977, 'W_D_1KI': 13.275772301284276, 'J_D_1KI': 4.6080431451871835}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 3433, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.349842309951782, "TIME_S_1KI": 3.01480987764398, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 866.6259407162667, "W": 65.83, "J_1KI": 252.43983126020004, "W_1KI": 19.175648121176813, "W_D": 31.058999999999997, "J_D": 408.8794636595249, "W_D_1KI": 9.047189047480337, "J_D_1KI": 2.635359466204584}

71
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/language.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 3.058311939239502}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.8774, 0.7244, 0.5547, ..., 0.2046, 0.1297, 0.0114])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 3.058311939239502 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '3433', '-m', 'matrices/389000+_cols/language.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.349842309951782}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.8074, 0.3012, 0.9549, ..., 0.2881, 0.3396, 0.4512])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.349842309951782 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.8074, 0.3012, 0.9549, ..., 0.2881, 0.3396, 0.4512])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.349842309951782 seconds
[39.97, 38.36, 38.86, 38.59, 38.26, 38.09, 38.21, 38.92, 38.29, 38.47]
[65.83]
13.164604902267456
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 3433, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.349842309951782, 'TIME_S_1KI': 3.01480987764398, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 866.6259407162667, 'W': 65.83}
[39.97, 38.36, 38.86, 38.59, 38.26, 38.09, 38.21, 38.92, 38.29, 38.47, 39.36, 38.28, 38.75, 39.51, 38.42, 38.7, 38.49, 38.22, 39.26, 38.62]
695.4200000000001
34.771
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 3433, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.349842309951782, 'TIME_S_1KI': 3.01480987764398, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 866.6259407162667, 'W': 65.83, 'J_1KI': 252.43983126020004, 'W_1KI': 19.175648121176813, 'W_D': 31.058999999999997, 'J_D': 408.8794636595249, 'W_D_1KI': 9.047189047480337, 'J_D_1KI': 2.635359466204584}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 1403, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.47270679473877, "TIME_S_1KI": 7.4645094759364, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1076.08722530365, "W": 76.53000000000002, "J_1KI": 766.9901819698147, "W_1KI": 54.54739843193158, "W_D": 40.18550000000002, "J_D": 565.0477354297641, "W_D_1KI": 28.642551674982194, "J_D_1KI": 20.415218585161934}

74
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,74 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 7.480671405792236}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.5771, 0.6006, 0.1014, ..., 0.3420, 0.9665, 0.9706])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 7.480671405792236 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1403', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 10.47270679473877}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.9737, 0.1599, 0.8628, ..., 0.5469, 0.5754, 0.2289])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.47270679473877 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.9737, 0.1599, 0.8628, ..., 0.5469, 0.5754, 0.2289])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 10.47270679473877 seconds
[39.8, 38.66, 38.76, 38.93, 38.51, 38.5, 38.48, 53.83, 45.0, 38.92]
[76.53]
14.060985565185547
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1403, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.47270679473877, 'TIME_S_1KI': 7.4645094759364, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1076.08722530365, 'W': 76.53000000000002}
[39.8, 38.66, 38.76, 38.93, 38.51, 38.5, 38.48, 53.83, 45.0, 38.92, 39.54, 38.42, 38.88, 38.56, 45.29, 38.95, 39.28, 38.53, 39.92, 38.52]
726.89
36.3445
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1403, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 10.47270679473877, 'TIME_S_1KI': 7.4645094759364, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1076.08722530365, 'W': 76.53000000000002, 'J_1KI': 766.9901819698147, 'W_1KI': 54.54739843193158, 'W_D': 40.18550000000002, 'J_D': 565.0477354297641, 'W_D_1KI': 28.642551674982194, 'J_D_1KI': 20.415218585161934}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 2778, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.518349885940552, "TIME_S_1KI": 3.786303054694223, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 942.0412521362306, "W": 70.4, "J_1KI": 339.1077221512709, "W_1KI": 25.341972642188626, "W_D": 35.57825000000001, "J_D": 476.0820906081797, "W_D_1KI": 12.807145428365734, "J_D_1KI": 4.61020353792863}

71
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 3.778977870941162}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.3385, 0.4156, 0.4762, ..., 0.6246, 0.7256, 0.2909])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 3.778977870941162 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2778', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.518349885940552}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.1706, 0.4199, 0.8169, ..., 0.9237, 0.2859, 0.4340])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.518349885940552 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.1706, 0.4199, 0.8169, ..., 0.9237, 0.2859, 0.4340])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.518349885940552 seconds
[39.45, 38.78, 38.96, 38.79, 38.45, 38.37, 38.35, 39.22, 38.56, 38.56]
[70.4]
13.381267786026001
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2778, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.518349885940552, 'TIME_S_1KI': 3.786303054694223, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 942.0412521362306, 'W': 70.4}
[39.45, 38.78, 38.96, 38.79, 38.45, 38.37, 38.35, 39.22, 38.56, 38.56, 40.19, 38.52, 38.44, 38.38, 38.77, 38.32, 38.58, 39.02, 38.41, 38.83]
696.435
34.821749999999994
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 2778, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.518349885940552, 'TIME_S_1KI': 3.786303054694223, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 942.0412521362306, 'W': 70.4, 'J_1KI': 339.1077221512709, 'W_1KI': 25.341972642188626, 'W_D': 35.57825000000001, 'J_D': 476.0820906081797, 'W_D_1KI': 12.807145428365734, 'J_D_1KI': 4.61020353792863}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 20.74174404144287, "TIME_S_1KI": 20.74174404144287, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1964.3499762821195, "W": 78.38, "J_1KI": 1964.3499762821195, "W_1KI": 78.38, "W_D": 43.477, "J_D": 1089.6152579588888, "W_D_1KI": 43.477, "J_D_1KI": 43.477}

51
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 20.74174404144287}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3044, 0.7914, 0.5459, ..., 0.2990, 0.3126, 0.6970])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 20.74174404144287 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.3044, 0.7914, 0.5459, ..., 0.2990, 0.3126, 0.6970])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 20.74174404144287 seconds
[39.18, 39.6, 38.95, 39.49, 38.43, 39.82, 38.88, 38.36, 38.96, 38.49]
[78.38]
25.061877727508545
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 20.74174404144287, 'TIME_S_1KI': 20.74174404144287, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1964.3499762821195, 'W': 78.38}
[39.18, 39.6, 38.95, 39.49, 38.43, 39.82, 38.88, 38.36, 38.96, 38.49, 39.91, 38.51, 38.34, 38.55, 38.51, 38.42, 38.42, 38.4, 38.39, 38.48]
698.06
34.903
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 20.74174404144287, 'TIME_S_1KI': 20.74174404144287, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1964.3499762821195, 'W': 78.38, 'J_1KI': 1964.3499762821195, 'W_1KI': 78.38, 'W_D': 43.477, 'J_D': 1089.6152579588888, 'W_D_1KI': 43.477, 'J_D_1KI': 43.477}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 12.424208641052246, "TIME_S_1KI": 12.424208641052246, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 879.7217792987824, "W": 52.38, "J_1KI": 879.7217792987824, "W_1KI": 52.38, "W_D": 36.156000000000006, "J_D": 607.2397986316682, "W_D_1KI": 36.156000000000006, "J_D_1KI": 36.156000000000006}

49
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/amazon0312.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 12.424208641052246}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6069, 0.8095, 0.9925, ..., 0.5957, 0.3239, 0.4137])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 12.424208641052246 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.6069, 0.8095, 0.9925, ..., 0.5957, 0.3239, 0.4137])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 12.424208641052246 seconds
[18.89, 17.73, 17.96, 17.88, 17.85, 17.69, 17.97, 17.83, 17.93, 18.01]
[52.38]
16.79499387741089
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 12.424208641052246, 'TIME_S_1KI': 12.424208641052246, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 879.7217792987824, 'W': 52.38}
[18.89, 17.73, 17.96, 17.88, 17.85, 17.69, 17.97, 17.83, 17.93, 18.01, 18.12, 18.25, 18.06, 18.22, 18.08, 17.89, 17.8, 17.75, 19.11, 17.94]
324.48
16.224
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 12.424208641052246, 'TIME_S_1KI': 12.424208641052246, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 879.7217792987824, 'W': 52.38, 'J_1KI': 879.7217792987824, 'W_1KI': 52.38, 'W_D': 36.156000000000006, 'J_D': 607.2397986316682, 'W_D_1KI': 36.156000000000006, 'J_D_1KI': 36.156000000000006}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1604, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.463838577270508, "TIME_S_1KI": 6.5235901354554295, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 792.353337585926, "W": 53.79, "J_1KI": 493.98587131292146, "W_1KI": 33.53491271820449, "W_D": 15.963500000000003, "J_D": 235.15026035606866, "W_D_1KI": 9.952306733167084, "J_D_1KI": 6.204680008208905}

71
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 6.544304132461548}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9984, 0.0550, 0.4152, ..., 0.8933, 0.3177, 0.3432])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 6.544304132461548 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1604', '-m', 'matrices/389000+_cols/darcy003.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.463838577270508}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.3553, 0.0914, 0.5617, ..., 0.2172, 0.2068, 0.5865])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.463838577270508 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.3553, 0.0914, 0.5617, ..., 0.2172, 0.2068, 0.5865])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.463838577270508 seconds
[51.54, 47.98, 24.21, 22.55, 39.25, 40.93, 45.39, 44.37, 40.99, 42.3]
[53.79]
14.73049521446228
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1604, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.463838577270508, 'TIME_S_1KI': 6.5235901354554295, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 792.353337585926, 'W': 53.79}
[51.54, 47.98, 24.21, 22.55, 39.25, 40.93, 45.39, 44.37, 40.99, 42.3, 51.25, 53.2, 44.13, 44.41, 42.29, 42.29, 43.1, 44.05, 43.73, 42.23]
756.53
37.826499999999996
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1604, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.463838577270508, 'TIME_S_1KI': 6.5235901354554295, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 792.353337585926, 'W': 53.79, 'J_1KI': 493.98587131292146, 'W_1KI': 33.53491271820449, 'W_D': 15.963500000000003, 'J_D': 235.15026035606866, 'W_D_1KI': 9.952306733167084, 'J_D_1KI': 6.204680008208905}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1567, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.482280731201172, "TIME_S_1KI": 6.6893942126363575, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 809.832340335846, "W": 53.88, "J_1KI": 516.8043014268321, "W_1KI": 34.38417358008934, "W_D": 37.52475, "J_D": 564.0080941540002, "W_D_1KI": 23.94687300574346, "J_D_1KI": 15.2819866022613}

74
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,74 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 6.698031663894653}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.6880, 0.8256, 0.6674, ..., 0.8572, 0.2017, 0.9423])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 6.698031663894653 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1567', '-m', 'matrices/389000+_cols/helm2d03.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 10.482280731201172}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9695, 0.5429, 0.1111, ..., 0.2474, 0.2323, 0.6789])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.482280731201172 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9695, 0.5429, 0.1111, ..., 0.2474, 0.2323, 0.6789])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 10.482280731201172 seconds
[18.37, 18.1, 22.31, 17.76, 18.12, 17.9, 17.88, 17.63, 17.79, 17.9]
[53.88]
15.030295848846436
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1567, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.482280731201172, 'TIME_S_1KI': 6.6893942126363575, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 809.832340335846, 'W': 53.88}
[18.37, 18.1, 22.31, 17.76, 18.12, 17.9, 17.88, 17.63, 17.79, 17.9, 18.29, 18.08, 17.89, 18.0, 17.96, 17.67, 18.29, 17.73, 17.87, 17.69]
327.105
16.35525
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1567, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 10.482280731201172, 'TIME_S_1KI': 6.6893942126363575, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 809.832340335846, 'W': 53.88, 'J_1KI': 516.8043014268321, 'W_1KI': 34.38417358008934, 'W_D': 37.52475, 'J_D': 564.0080941540002, 'W_D_1KI': 23.94687300574346, 'J_D_1KI': 15.2819866022613}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1711, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.39021348953247, "TIME_S_1KI": 6.072597013169182, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 758.5677815818786, "W": 52.04, "J_1KI": 443.34762219864325, "W_1KI": 30.414962010520163, "W_D": 35.8595, "J_D": 522.710633428812, "W_D_1KI": 20.95821157218001, "J_D_1KI": 12.249100860420812}

71
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/language.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 6.135177135467529}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.0547, 0.0947, 0.9321, ..., 0.9094, 0.0107, 0.8738])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 6.135177135467529 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1711', '-m', 'matrices/389000+_cols/language.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 10.39021348953247}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.6001, 0.7097, 0.4908, ..., 0.7271, 0.7976, 0.2970])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.39021348953247 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.6001, 0.7097, 0.4908, ..., 0.7271, 0.7976, 0.2970])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 10.39021348953247 seconds
[18.03, 17.94, 17.82, 17.99, 17.74, 17.56, 17.76, 17.68, 17.79, 18.07]
[52.04]
14.576629161834717
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1711, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.39021348953247, 'TIME_S_1KI': 6.072597013169182, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 758.5677815818786, 'W': 52.04}
[18.03, 17.94, 17.82, 17.99, 17.74, 17.56, 17.76, 17.68, 17.79, 18.07, 18.72, 17.93, 17.72, 18.0, 18.39, 17.97, 18.72, 17.86, 18.44, 17.78]
323.60999999999996
16.1805
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1711, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 10.39021348953247, 'TIME_S_1KI': 6.072597013169182, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 758.5677815818786, 'W': 52.04, 'J_1KI': 443.34762219864325, 'W_1KI': 30.414962010520163, 'W_D': 35.8595, 'J_D': 522.710633428812, 'W_D_1KI': 20.95821157218001, 'J_D_1KI': 12.249100860420812}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 14.3206205368042, "TIME_S_1KI": 14.3206205368042, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1077.2933460760116, "W": 53.62, "J_1KI": 1077.2933460760116, "W_1KI": 53.62, "W_D": 37.37875, "J_D": 750.9861741819977, "W_D_1KI": 37.37875, "J_D_1KI": 37.37875}

51
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/marine1.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 14.3206205368042}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.4405, 0.4136, 0.9296, ..., 0.1477, 0.1453, 0.8762])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 14.3206205368042 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.4405, 0.4136, 0.9296, ..., 0.1477, 0.1453, 0.8762])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 14.3206205368042 seconds
[18.25, 17.59, 17.7, 17.8, 18.1, 17.82, 18.62, 21.23, 18.03, 17.76]
[53.62]
20.091259717941284
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 14.3206205368042, 'TIME_S_1KI': 14.3206205368042, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1077.2933460760116, 'W': 53.62}
[18.25, 17.59, 17.7, 17.8, 18.1, 17.82, 18.62, 21.23, 18.03, 17.76, 18.27, 17.65, 17.53, 17.72, 17.84, 17.65, 17.87, 17.77, 17.96, 17.61]
324.825
16.24125
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 14.3206205368042, 'TIME_S_1KI': 14.3206205368042, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1077.2933460760116, 'W': 53.62, 'J_1KI': 1077.2933460760116, 'W_1KI': 53.62, 'W_D': 37.37875, 'J_D': 750.9861741819977, 'W_D_1KI': 37.37875, 'J_D_1KI': 37.37875}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1598, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.48720407485962, "TIME_S_1KI": 6.562705929198761, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 768.8178126811981, "W": 53.0, "J_1KI": 481.1125235802241, "W_1KI": 33.16645807259074, "W_D": 36.777, "J_D": 533.4870320184231, "W_D_1KI": 23.01439299123905, "J_D_1KI": 14.40199811717087}

71
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,71 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 6.568377256393433}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.1410, 0.8504, 0.4141, ..., 0.6370, 0.5152, 0.1646])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 6.568377256393433 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1598', '-m', 'matrices/389000+_cols/mario002.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 10.48720407485962}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9888, 0.3844, 0.2800, ..., 0.8268, 0.5179, 0.1169])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.48720407485962 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.9888, 0.3844, 0.2800, ..., 0.8268, 0.5179, 0.1169])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 10.48720407485962 seconds
[18.78, 17.92, 18.04, 18.03, 17.82, 17.72, 18.29, 17.74, 17.92, 17.72]
[53.0]
14.505996465682983
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1598, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.48720407485962, 'TIME_S_1KI': 6.562705929198761, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 768.8178126811981, 'W': 53.0}
[18.78, 17.92, 18.04, 18.03, 17.82, 17.72, 18.29, 17.74, 17.92, 17.72, 18.42, 18.18, 17.97, 17.99, 17.82, 17.72, 17.94, 18.12, 18.4, 18.76]
324.4599999999999
16.222999999999995
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1598, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 10.48720407485962, 'TIME_S_1KI': 6.562705929198761, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 768.8178126811981, 'W': 53.0, 'J_1KI': 481.1125235802241, 'W_1KI': 33.16645807259074, 'W_D': 36.777, 'J_D': 533.4870320184231, 'W_D_1KI': 23.01439299123905, 'J_D_1KI': 14.40199811717087}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 37.58896088600159, "TIME_S_1KI": 37.58896088600159, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 2465.35528427124, "W": 53.12, "J_1KI": 2465.35528427124, "W_1KI": 53.12, "W_D": 36.98599999999999, "J_D": 1716.5593099408145, "W_D_1KI": 36.98599999999999, "J_D_1KI": 36.98599999999999}

51
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1000', '-m', 'matrices/389000+_cols/test1.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 37.58896088600159}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.9581, 0.4048, 0.0262, ..., 0.9819, 0.7450, 0.5527])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 37.58896088600159 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.9581, 0.4048, 0.0262, ..., 0.9819, 0.7450, 0.5527])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 37.58896088600159 seconds
[18.09, 17.76, 17.72, 17.96, 17.74, 17.95, 17.93, 17.84, 17.96, 18.58]
[53.12]
46.41105580329895
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 37.58896088600159, 'TIME_S_1KI': 37.58896088600159, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2465.35528427124, 'W': 53.12}
[18.09, 17.76, 17.72, 17.96, 17.74, 17.95, 17.93, 17.84, 17.96, 18.58, 18.32, 17.63, 17.7, 18.87, 17.96, 17.88, 17.96, 17.6, 17.83, 17.79]
322.68000000000006
16.134000000000004
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 37.58896088600159, 'TIME_S_1KI': 37.58896088600159, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2465.35528427124, 'W': 53.12, 'J_1KI': 2465.35528427124, 'W_1KI': 53.12, 'W_D': 36.98599999999999, 'J_D': 1716.5593099408145, 'W_D_1KI': 36.98599999999999, 'J_D_1KI': 36.98599999999999}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 30.323144912719727, "TIME_S_1KI": 30.323144912719727, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 2708.072883968354, "W": 77.40203729318671, "J_1KI": 2708.072883968354, "W_1KI": 77.40203729318671, "W_D": 53.61603729318671, "J_D": 1875.8697033972746, "W_D_1KI": 53.61603729318671, "J_D_1KI": 53.61603729318671}

49
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_amazon0312.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/amazon0312.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 30.323144912719727}
/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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.9519, 0.7886, 0.4122, ..., 0.0191, 0.4041, 0.8787])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 30.323144912719727 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, 5, 10, ..., 3200428,
3200438, 3200440]),
col_indices=tensor([ 1, 2, 3, ..., 400724, 6009,
400707]),
values=tensor([1., 1., 1., ..., 1., 1., 1.]), size=(400727, 400727),
nnz=3200440, layout=torch.sparse_csr)
tensor([0.9519, 0.7886, 0.4122, ..., 0.0191, 0.4041, 0.8787])
Matrix Type: SuiteSparse
Matrix: amazon0312
Matrix Format: csr
Shape: torch.Size([400727, 400727])
Rows: 400727
Size: 160582128529
NNZ: 3200440
Density: 1.9930237750099465e-05
Time: 30.323144912719727 seconds
[26.56, 26.36, 26.4, 26.4, 26.28, 26.16, 26.4, 26.4, 26.6, 26.56]
[26.56, 26.64, 26.44, 29.44, 30.6, 36.2, 50.64, 63.2, 76.56, 90.36, 96.08, 94.84, 94.28, 93.0, 94.08, 94.08, 94.76, 95.56, 97.52, 96.08, 95.76, 91.48, 89.72, 90.12, 89.92, 89.6, 90.64, 91.72, 90.48, 91.12, 89.6, 89.6, 90.12, 90.48]
34.98710083961487
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 30.323144912719727, 'TIME_S_1KI': 30.323144912719727, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2708.072883968354, 'W': 77.40203729318671}
[26.56, 26.36, 26.4, 26.4, 26.28, 26.16, 26.4, 26.4, 26.6, 26.56, 26.8, 26.84, 26.76, 26.72, 26.52, 26.44, 26.32, 26.16, 26.0, 26.0]
475.72
23.786
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'amazon0312', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400727, 400727], 'MATRIX_ROWS': 400727, 'MATRIX_SIZE': 160582128529, 'MATRIX_NNZ': 3200440, 'MATRIX_DENSITY': 1.9930237750099465e-05, 'TIME_S': 30.323144912719727, 'TIME_S_1KI': 30.323144912719727, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2708.072883968354, 'W': 77.40203729318671, 'J_1KI': 2708.072883968354, 'W_1KI': 77.40203729318671, 'W_D': 53.61603729318671, 'J_D': 1875.8697033972746, 'W_D_1KI': 53.61603729318671, 'J_D_1KI': 53.61603729318671}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_darcy003.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 20.392087936401367, "TIME_S_1KI": 20.392087936401367, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1566.7498482894898, "W": 69.10901709242208, "J_1KI": 1566.7498482894898, "W_1KI": 69.10901709242208, "W_D": 45.48401709242208, "J_D": 1031.1545421612263, "W_D_1KI": 45.48401709242208, "J_D_1KI": 45.48401709242208}

49
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_darcy003.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/darcy003.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "darcy003", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 20.392087936401367}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.5706, 0.0924, 0.8150, ..., 0.7995, 0.0048, 0.8110])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 20.392087936401367 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.5706, 0.0924, 0.8150, ..., 0.7995, 0.0048, 0.8110])
Matrix Type: SuiteSparse
Matrix: darcy003
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 20.392087936401367 seconds
[26.52, 26.44, 26.28, 26.32, 26.32, 26.12, 26.16, 26.12, 26.12, 26.08]
[26.52, 26.68, 29.52, 30.56, 30.56, 36.56, 50.04, 66.72, 78.32, 93.24, 95.96, 93.96, 92.08, 92.12, 90.56, 91.2, 90.64, 89.92, 88.64, 89.44, 89.44, 91.68]
22.670700788497925
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 20.392087936401367, 'TIME_S_1KI': 20.392087936401367, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1566.7498482894898, 'W': 69.10901709242208}
[26.52, 26.44, 26.28, 26.32, 26.32, 26.12, 26.16, 26.12, 26.12, 26.08, 26.48, 26.28, 26.28, 26.4, 26.2, 26.0, 26.28, 26.28, 26.28, 26.16]
472.5
23.625
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'darcy003', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 20.392087936401367, 'TIME_S_1KI': 20.392087936401367, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1566.7498482894898, 'W': 69.10901709242208, 'J_1KI': 1566.7498482894898, 'W_1KI': 69.10901709242208, 'W_D': 45.48401709242208, 'J_D': 1031.1545421612263, 'W_D_1KI': 45.48401709242208, 'J_D_1KI': 45.48401709242208}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_helm2d03.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 21.815975189208984, "TIME_S_1KI": 21.815975189208984, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1916.1918925571438, "W": 74.64919082171829, "J_1KI": 1916.1918925571438, "W_1KI": 74.64919082171829, "W_D": 51.65619082171829, "J_D": 1325.9778567373746, "W_D_1KI": 51.65619082171829, "J_D_1KI": 51.65619082171829}

51
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_helm2d03.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/helm2d03.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "helm2d03", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392257, 392257], "MATRIX_ROWS": 392257, "MATRIX_SIZE": 153865554049, "MATRIX_NNZ": 2741935, "MATRIX_DENSITY": 1.7820330332848923e-05, "TIME_S": 21.815975189208984}
/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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9058, 0.4925, 0.9859, ..., 0.1438, 0.2004, 0.4986])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 21.815975189208984 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, 7, 14, ..., 2741921,
2741928, 2741935]),
col_indices=tensor([ 0, 98273, 133833, ..., 392252, 392254,
392256]),
values=tensor([ 3.4808, -0.6217, -0.5806, ..., -0.6940, -0.7602,
3.5476]), size=(392257, 392257), nnz=2741935,
layout=torch.sparse_csr)
tensor([0.9058, 0.4925, 0.9859, ..., 0.1438, 0.2004, 0.4986])
Matrix Type: SuiteSparse
Matrix: helm2d03
Matrix Format: csr
Shape: torch.Size([392257, 392257])
Rows: 392257
Size: 153865554049
NNZ: 2741935
Density: 1.7820330332848923e-05
Time: 21.815975189208984 seconds
[25.36, 25.4, 25.56, 25.56, 25.6, 25.52, 25.52, 25.28, 25.36, 25.16]
[25.0, 24.96, 25.32, 27.72, 29.24, 40.8, 57.48, 70.72, 86.32, 97.72, 97.72, 96.24, 97.0, 95.72, 95.32, 93.32, 95.8, 96.36, 95.36, 95.72, 94.88, 94.16, 93.92, 92.68, 93.4]
25.669292211532593
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 21.815975189208984, 'TIME_S_1KI': 21.815975189208984, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1916.1918925571438, 'W': 74.64919082171829}
[25.36, 25.4, 25.56, 25.56, 25.6, 25.52, 25.52, 25.28, 25.36, 25.16, 25.44, 25.44, 25.28, 25.52, 25.88, 25.84, 25.68, 25.88, 25.8, 25.52]
459.86
22.993000000000002
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'helm2d03', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392257, 392257], 'MATRIX_ROWS': 392257, 'MATRIX_SIZE': 153865554049, 'MATRIX_NNZ': 2741935, 'MATRIX_DENSITY': 1.7820330332848923e-05, 'TIME_S': 21.815975189208984, 'TIME_S_1KI': 21.815975189208984, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1916.1918925571438, 'W': 74.64919082171829, 'J_1KI': 1916.1918925571438, 'W_1KI': 74.64919082171829, 'W_D': 51.65619082171829, 'J_D': 1325.9778567373746, 'W_D_1KI': 51.65619082171829, 'J_D_1KI': 51.65619082171829}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_language.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 11.040250062942505, "TIME_S_1KI": 11.040250062942505, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1365.697071151733, "W": 66.06671490426629, "J_1KI": 1365.697071151733, "W_1KI": 66.06671490426629, "W_D": 44.15271490426629, "J_D": 912.7021604681013, "W_D_1KI": 44.15271490426629, "J_D_1KI": 44.15271490426629}

49
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_language.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/language.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "language", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [399130, 399130], "MATRIX_ROWS": 399130, "MATRIX_SIZE": 159304756900, "MATRIX_NNZ": 1216334, "MATRIX_DENSITY": 7.635264782228233e-06, "TIME_S": 11.040250062942505}
/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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.4838, 0.0445, 0.9105, ..., 0.8272, 0.1700, 0.2253])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 11.040250062942505 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, 3, ..., 1216330,
1216332, 1216334]),
col_indices=tensor([ 0, 0, 1, ..., 399128, 399125,
399129]),
values=tensor([ 1., -1., 1., ..., 1., -1., 1.]),
size=(399130, 399130), nnz=1216334, layout=torch.sparse_csr)
tensor([0.4838, 0.0445, 0.9105, ..., 0.8272, 0.1700, 0.2253])
Matrix Type: SuiteSparse
Matrix: language
Matrix Format: csr
Shape: torch.Size([399130, 399130])
Rows: 399130
Size: 159304756900
NNZ: 1216334
Density: 7.635264782228233e-06
Time: 11.040250062942505 seconds
[24.56, 24.52, 24.08, 24.04, 23.92, 23.84, 23.84, 23.84, 24.04, 24.12]
[24.28, 24.56, 24.56, 28.8, 30.88, 44.64, 59.0, 72.44, 83.72, 90.96, 88.6, 88.6, 85.88, 86.24, 86.16, 87.16, 86.96, 88.8, 89.32, 90.72]
20.67148447036743
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 11.040250062942505, 'TIME_S_1KI': 11.040250062942505, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1365.697071151733, 'W': 66.06671490426629}
[24.56, 24.52, 24.08, 24.04, 23.92, 23.84, 23.84, 23.84, 24.04, 24.12, 24.64, 24.56, 24.68, 24.52, 24.56, 24.64, 24.8, 24.72, 24.72, 24.6]
438.28
21.913999999999998
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'language', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [399130, 399130], 'MATRIX_ROWS': 399130, 'MATRIX_SIZE': 159304756900, 'MATRIX_NNZ': 1216334, 'MATRIX_DENSITY': 7.635264782228233e-06, 'TIME_S': 11.040250062942505, 'TIME_S_1KI': 11.040250062942505, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1365.697071151733, 'W': 66.06671490426629, 'J_1KI': 1365.697071151733, 'W_1KI': 66.06671490426629, 'W_D': 44.15271490426629, 'J_D': 912.7021604681013, 'W_D_1KI': 44.15271490426629, 'J_D_1KI': 44.15271490426629}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_marine1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 47.31629490852356, "TIME_S_1KI": 47.31629490852356, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 3588.3839783859257, "W": 77.04772778103, "J_1KI": 3588.3839783859257, "W_1KI": 77.04772778103, "W_D": 53.199727781030006, "J_D": 2477.698646305085, "W_D_1KI": 53.199727781030006, "J_D_1KI": 53.199727781030006}

51
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_marine1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/marine1.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "marine1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400320, 400320], "MATRIX_ROWS": 400320, "MATRIX_SIZE": 160256102400, "MATRIX_NNZ": 6226538, "MATRIX_DENSITY": 3.885367175883594e-05, "TIME_S": 47.31629490852356}
/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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.9557, 0.7622, 0.1453, ..., 0.2002, 0.7167, 0.8732])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 47.31629490852356 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, 7, 18, ..., 6226522,
6226531, 6226538]),
col_indices=tensor([ 0, 1, 10383, ..., 400315, 400318,
400319]),
values=tensor([ 6.2373e+03, -1.8964e+00, -5.7529e+00, ...,
-6.8099e-01, -6.4187e-01, 1.7595e+01]),
size=(400320, 400320), nnz=6226538, layout=torch.sparse_csr)
tensor([0.9557, 0.7622, 0.1453, ..., 0.2002, 0.7167, 0.8732])
Matrix Type: SuiteSparse
Matrix: marine1
Matrix Format: csr
Shape: torch.Size([400320, 400320])
Rows: 400320
Size: 160256102400
NNZ: 6226538
Density: 3.885367175883594e-05
Time: 47.31629490852356 seconds
[26.68, 26.68, 26.56, 26.48, 26.24, 26.08, 26.16, 25.8, 26.04, 26.04]
[26.4, 26.6, 26.72, 27.8, 29.68, 36.68, 47.12, 59.56, 71.44, 82.04, 90.0, 90.2, 89.8, 91.64, 91.36, 89.28, 89.28, 88.72, 90.36, 90.56, 88.08, 89.32, 91.24, 89.92, 91.16, 94.4, 94.0, 92.04, 91.28, 91.0, 89.92, 89.76, 89.76, 89.8, 90.52, 90.56, 92.08, 91.72, 90.16, 89.0, 89.52, 88.96, 88.76, 89.0, 87.92]
46.57352113723755
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 47.31629490852356, 'TIME_S_1KI': 47.31629490852356, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 3588.3839783859257, 'W': 77.04772778103}
[26.68, 26.68, 26.56, 26.48, 26.24, 26.08, 26.16, 25.8, 26.04, 26.04, 26.76, 26.48, 26.44, 26.48, 26.72, 26.72, 26.92, 26.92, 27.0, 27.0]
476.9599999999999
23.847999999999995
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'marine1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [400320, 400320], 'MATRIX_ROWS': 400320, 'MATRIX_SIZE': 160256102400, 'MATRIX_NNZ': 6226538, 'MATRIX_DENSITY': 3.885367175883594e-05, 'TIME_S': 47.31629490852356, 'TIME_S_1KI': 47.31629490852356, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 3588.3839783859257, 'W': 77.04772778103, 'J_1KI': 3588.3839783859257, 'W_1KI': 77.04772778103, 'W_D': 53.199727781030006, 'J_D': 2477.698646305085, 'W_D_1KI': 53.199727781030006, 'J_D_1KI': 53.199727781030006}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_mario002.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 26.970608472824097, "TIME_S_1KI": 26.970608472824097, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1533.6805934524534, "W": 70.7905728448904, "J_1KI": 1533.6805934524534, "W_1KI": 70.7905728448904, "W_D": 46.71357284489041, "J_D": 1012.0514249830246, "W_D_1KI": 46.71357284489041, "J_D_1KI": 46.71357284489041}

49
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_mario002.output Normal file
View File

@ -0,0 +1,49 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/mario002.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "mario002", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [389874, 389874], "MATRIX_ROWS": 389874, "MATRIX_SIZE": 152001735876, "MATRIX_NNZ": 2101242, "MATRIX_DENSITY": 1.3823802655215408e-05, "TIME_S": 26.970608472824097}
/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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0717, 0.4634, 0.0880, ..., 0.8346, 0.7497, 0.2295])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 26.970608472824097 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, 7, ..., 2101236,
2101239, 2101242]),
col_indices=tensor([ 0, 1027, 1028, ..., 196606, 233926,
234127]),
values=tensor([ 1., 0., 0., ..., -1., -1., -1.]),
size=(389874, 389874), nnz=2101242, layout=torch.sparse_csr)
tensor([0.0717, 0.4634, 0.0880, ..., 0.8346, 0.7497, 0.2295])
Matrix Type: SuiteSparse
Matrix: mario002
Matrix Format: csr
Shape: torch.Size([389874, 389874])
Rows: 389874
Size: 152001735876
NNZ: 2101242
Density: 1.3823802655215408e-05
Time: 26.970608472824097 seconds
[27.04, 26.84, 26.96, 26.8, 26.8, 27.0, 26.88, 26.6, 26.44, 26.44]
[26.12, 26.4, 26.44, 27.4, 29.4, 37.08, 54.84, 69.0, 83.6, 95.56, 98.68, 97.52, 97.96, 97.76, 97.76, 97.8, 98.4, 95.2, 93.56, 91.64, 89.8]
21.665040016174316
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 26.970608472824097, 'TIME_S_1KI': 26.970608472824097, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1533.6805934524534, 'W': 70.7905728448904}
[27.04, 26.84, 26.96, 26.8, 26.8, 27.0, 26.88, 26.6, 26.44, 26.44, 26.88, 26.64, 26.8, 27.08, 27.0, 26.92, 26.68, 26.44, 26.32, 26.32]
481.53999999999996
24.076999999999998
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'mario002', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [389874, 389874], 'MATRIX_ROWS': 389874, 'MATRIX_SIZE': 152001735876, 'MATRIX_NNZ': 2101242, 'MATRIX_DENSITY': 1.3823802655215408e-05, 'TIME_S': 26.970608472824097, 'TIME_S_1KI': 26.970608472824097, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1533.6805934524534, 'W': 70.7905728448904, 'J_1KI': 1533.6805934524534, 'W_1KI': 70.7905728448904, 'W_D': 46.71357284489041, 'J_D': 1012.0514249830246, 'W_D_1KI': 46.71357284489041, 'J_D_1KI': 46.71357284489041}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_test1.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 1000, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 92.92496466636658, "TIME_S_1KI": 92.92496466636658, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 8599.791397418978, "W": 85.10134779265783, "J_1KI": 8599.791397418978, "W_1KI": 85.10134779265783, "W_D": 60.408347792657835, "J_D": 6104.4766405498995, "W_D_1KI": 60.408347792657835, "J_D_1KI": 60.408347792657835}

51
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_test1.output Normal file
View File

@ -0,0 +1,51 @@
['apptainer', 'run', 'pytorch-altra.sif', '-c', 'numactl --cpunodebind=0 --membind=0 python3 spmv.py suitesparse csr 1000 -m matrices/389000+_cols/test1.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "test1", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [392908, 392908], "MATRIX_ROWS": 392908, "MATRIX_SIZE": 154376696464, "MATRIX_NNZ": 12968200, "MATRIX_DENSITY": 8.400361127706946e-05, "TIME_S": 92.92496466636658}
/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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.1407, 0.4697, 0.2203, ..., 0.3353, 0.2584, 0.9591])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 92.92496466636658 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, 24, 48, ..., 12968181,
12968191, 12968200]),
col_indices=tensor([ 0, 1, 8, ..., 392905, 392906,
392907]),
values=tensor([1.0000e+00, 0.0000e+00, 0.0000e+00, ...,
0.0000e+00, 0.0000e+00, 2.1156e-17]),
size=(392908, 392908), nnz=12968200, layout=torch.sparse_csr)
tensor([0.1407, 0.4697, 0.2203, ..., 0.3353, 0.2584, 0.9591])
Matrix Type: SuiteSparse
Matrix: test1
Matrix Format: csr
Shape: torch.Size([392908, 392908])
Rows: 392908
Size: 154376696464
NNZ: 12968200
Density: 8.400361127706946e-05
Time: 92.92496466636658 seconds
[27.6, 27.24, 27.2, 27.24, 27.04, 26.92, 26.76, 26.64, 26.72, 26.64]
[26.8, 26.88, 30.16, 31.04, 31.04, 35.8, 41.52, 47.96, 57.12, 73.0, 77.8, 92.16, 95.08, 96.56, 94.84, 94.48, 94.88, 92.8, 91.28, 91.28, 91.72, 91.6, 89.08, 89.24, 90.12, 89.76, 89.48, 89.84, 90.24, 88.68, 88.6, 90.24, 88.36, 90.44, 90.44, 93.28, 93.6, 95.92, 95.52, 91.44, 91.84, 89.96, 91.0, 91.04, 90.84, 90.68, 88.92, 93.0, 95.6, 97.32, 97.32, 98.32, 96.56, 92.48, 91.04, 92.0, 91.72, 94.24, 92.8, 92.52, 91.64, 93.04, 93.56, 98.0, 99.0, 97.0, 97.0, 94.36, 93.2, 93.0, 91.36, 92.44, 95.16, 96.28, 96.76, 96.4, 97.92, 92.76, 93.72, 94.64, 95.76, 96.08, 96.08, 99.8, 99.2, 97.52, 97.16, 96.92, 97.2, 99.08, 99.0, 97.56, 97.8, 97.16, 97.4, 98.52, 99.92, 98.48]
101.05352759361267
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 92.92496466636658, 'TIME_S_1KI': 92.92496466636658, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 8599.791397418978, 'W': 85.10134779265783}
[27.6, 27.24, 27.2, 27.24, 27.04, 26.92, 26.76, 26.64, 26.72, 26.64, 27.96, 27.76, 27.8, 27.8, 27.96, 28.08, 28.04, 27.96, 27.76, 27.68]
493.86
24.693
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 1000, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'test1', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [392908, 392908], 'MATRIX_ROWS': 392908, 'MATRIX_SIZE': 154376696464, 'MATRIX_NNZ': 12968200, 'MATRIX_DENSITY': 8.400361127706946e-05, 'TIME_S': 92.92496466636658, 'TIME_S_1KI': 92.92496466636658, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 8599.791397418978, 'W': 85.10134779265783, 'J_1KI': 8599.791397418978, 'W_1KI': 85.10134779265783, 'W_D': 60.408347792657835, 'J_D': 6104.4766405498995, 'W_D_1KI': 60.408347792657835, 'J_D_1KI': 60.408347792657835}

1
pytorch/output_389000+_maxcore/epyc_7313p_max_csr_10_10_10_amazon0312.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 19947, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "amazon0312", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [400727, 400727], "MATRIX_ROWS": 400727, "MATRIX_SIZE": 160582128529, "MATRIX_NNZ": 3200440, "MATRIX_DENSITY": 1.9930237750099465e-05, "TIME_S": 10.311090230941772, "TIME_S_1KI": 0.5169243611040143, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1933.8912408947945, "W": 148.79, "J_1KI": 96.951483475951, "W_1KI": 7.459267057702912, "W_D": 113.0615, "J_D": 1469.5116911917924, "W_D_1KI": 5.668095452950318, "J_D_1KI": 0.28415779079311765}

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