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

More 389000+ data

Browse Source
This commit is contained in:
cephi 2024-12-17 20:02:52 -05:00
parent 0def30ded2
commit 346d7162df
18 changed files with 721 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
pytorch/output_389000+_16core/altra_16_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 16, "ITERATIONS": 276, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.460163831710815, "TIME_S_1KI": 37.89914431779281, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 556.1444871425629, "W": 33.16277212022306, "J_1KI": 2015.016257762909, "W_1KI": 120.15497145008355, "W_D": 17.98077212022306, "J_D": 301.54015029191976, "W_D_1KI": 65.14772507327196, "J_D_1KI": 236.04248214953608}

74
pytorch/output_389000+_16core/altra_16_csr_10_10_10_msdoor.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 100 -m matrices/389000+_cols/msdoor.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 3.7967092990875244}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.9111, 0.9233, 0.5316, ..., 0.5920, 0.1518, 0.8793])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 3.7967092990875244 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 276 -m matrices/389000+_cols/msdoor.mtx -c 16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.460163831710815}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.6477, 0.4686, 0.6140, ..., 0.5698, 0.0750, 0.2417])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.460163831710815 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.6477, 0.4686, 0.6140, ..., 0.5698, 0.0750, 0.2417])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.460163831710815 seconds
[16.64, 16.68, 16.6, 16.4, 16.68, 16.76, 16.96, 17.0, 17.04, 17.04]
[16.68, 16.64, 16.8, 20.28, 21.28, 25.56, 26.76, 29.96, 35.88, 41.68, 46.2, 50.76, 50.72, 50.92, 50.4, 50.36]
16.77014470100403
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 276, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.460163831710815, 'TIME_S_1KI': 37.89914431779281, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 556.1444871425629, 'W': 33.16277212022306}
[16.64, 16.68, 16.6, 16.4, 16.68, 16.76, 16.96, 17.0, 17.04, 17.04, 16.92, 17.2, 17.16, 17.16, 17.04, 16.8, 17.0, 16.76, 16.76, 16.68]
303.64
15.181999999999999
{'CPU': 'Altra', 'CORES': 16, 'ITERATIONS': 276, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.460163831710815, 'TIME_S_1KI': 37.89914431779281, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 556.1444871425629, 'W': 33.16277212022306, 'J_1KI': 2015.016257762909, 'W_1KI': 120.15497145008355, 'W_D': 17.98077212022306, 'J_D': 301.54015029191976, 'W_D_1KI': 65.14772507327196, 'J_D_1KI': 236.04248214953608}

1
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 2078, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.368424892425537, "TIME_S_1KI": 4.989617368828458, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1768.3314125061036, "W": 128.56, "J_1KI": 850.9775806092895, "W_1KI": 61.86717998075073, "W_D": 92.98825, "J_D": 1279.0451421046257, "W_D_1KI": 44.74891722810394, "J_D_1KI": 21.534608868192468}

74
pytorch/output_389000+_16core/epyc_7313p_16_csr_10_10_10_msdoor.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', '100', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 0.505284309387207}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0979, 0.4410, 0.6794, ..., 0.5232, 0.5066, 0.8230])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 0.505284309387207 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2078', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.368424892425537}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.1753, 0.8799, 0.4934, ..., 0.3593, 0.7648, 0.6193])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.368424892425537 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.1753, 0.8799, 0.4934, ..., 0.3593, 0.7648, 0.6193])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.368424892425537 seconds
[41.45, 39.25, 39.16, 39.53, 39.32, 39.64, 39.36, 39.48, 39.45, 39.21]
[128.56]
13.754911422729492
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2078, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.368424892425537, 'TIME_S_1KI': 4.989617368828458, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1768.3314125061036, 'W': 128.56}
[41.45, 39.25, 39.16, 39.53, 39.32, 39.64, 39.36, 39.48, 39.45, 39.21, 39.72, 39.17, 39.05, 38.92, 39.31, 39.68, 39.47, 38.94, 39.39, 44.25]
711.4350000000001
35.57175
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2078, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.368424892425537, 'TIME_S_1KI': 4.989617368828458, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1768.3314125061036, 'W': 128.56, 'J_1KI': 850.9775806092895, 'W_1KI': 61.86717998075073, 'W_D': 92.98825, 'J_D': 1279.0451421046257, 'W_D_1KI': 44.74891722810394, 'J_D_1KI': 21.534608868192468}

1
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 1333, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.536967515945435, "TIME_S_1KI": 7.9047018124121795, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1980.6726791381834, "W": 66.6, "J_1KI": 1485.8759783482246, "W_1KI": 49.962490622655665, "W_D": 49.48649999999999, "J_D": 1471.7200981407163, "W_D_1KI": 37.12415603900975, "J_D_1KI": 27.8500795491446}

97
pytorch/output_389000+_16core/xeon_4216_16_csr_10_10_10_msdoor.output Normal file
View File

@ -0,0 +1,97 @@
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:16}', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 1.0501856803894043}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.5509, 0.1350, 0.4677, ..., 0.8467, 0.1931, 0.8337])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 1.0501856803894043 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', '999', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 7.86853814125061}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0083, 0.1850, 0.6844, ..., 0.3666, 0.9050, 0.1703])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 7.86853814125061 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', '1333', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '16']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.536967515945435}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.5281, 0.2288, 0.8014, ..., 0.5844, 0.6890, 0.8626])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.536967515945435 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.5281, 0.2288, 0.8014, ..., 0.5844, 0.6890, 0.8626])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.536967515945435 seconds
[19.73, 18.62, 18.78, 18.65, 18.85, 18.62, 18.87, 18.75, 18.73, 18.48]
[66.6]
29.739830017089844
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1333, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.536967515945435, 'TIME_S_1KI': 7.9047018124121795, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1980.6726791381834, 'W': 66.6}
[19.73, 18.62, 18.78, 18.65, 18.85, 18.62, 18.87, 18.75, 18.73, 18.48, 20.53, 18.58, 18.75, 19.13, 18.65, 18.58, 22.49, 18.38, 18.94, 19.06]
342.27
17.1135
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1333, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.536967515945435, 'TIME_S_1KI': 7.9047018124121795, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1980.6726791381834, 'W': 66.6, 'J_1KI': 1485.8759783482246, 'W_1KI': 49.962490622655665, 'W_D': 49.48649999999999, 'J_D': 1471.7200981407163, 'W_D_1KI': 37.12415603900975, 'J_D_1KI': 27.8500795491446}

1
pytorch/output_389000+_1core/altra_1_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 1, "ITERATIONS": 100, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 43.16092610359192, "TIME_S_1KI": 431.6092610359192, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1189.33110124588, "W": 22.848960166729626, "J_1KI": 11893.3110124588, "W_1KI": 228.48960166729626, "W_D": 4.571960166729628, "J_D": 237.97907564592364, "W_D_1KI": 45.71960166729628, "J_D_1KI": 457.19601667296286}

51
pytorch/output_389000+_1core/altra_1_csr_10_10_10_msdoor.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 100 -m matrices/389000+_cols/msdoor.mtx -c 1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 43.16092610359192}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0523, 0.1527, 0.8721, ..., 0.0519, 0.6042, 0.4109])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 43.16092610359192 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0523, 0.1527, 0.8721, ..., 0.0519, 0.6042, 0.4109])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 43.16092610359192 seconds
[20.12, 19.92, 20.36, 20.44, 20.2, 20.04, 20.08, 19.84, 20.04, 20.04]
[20.2, 20.44, 20.76, 24.08, 25.44, 27.84, 28.48, 28.32, 27.16, 26.68, 25.2, 25.2, 24.44, 24.64, 24.56, 24.44, 24.4, 24.6, 24.4, 24.4, 24.36, 24.32, 24.28, 24.44, 24.56, 24.44, 24.44, 24.56, 24.52, 24.72, 25.24, 25.12, 25.36, 25.36, 25.56, 25.16, 25.12, 25.12, 24.84, 24.48, 24.44, 24.16, 24.12, 24.36, 24.32, 24.44, 24.28, 24.32, 24.12]
52.05186986923218
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 100, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 43.16092610359192, 'TIME_S_1KI': 431.6092610359192, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1189.33110124588, 'W': 22.848960166729626}
[20.12, 19.92, 20.36, 20.44, 20.2, 20.04, 20.08, 19.84, 20.04, 20.04, 20.28, 20.36, 20.48, 20.44, 20.44, 20.6, 20.72, 20.64, 20.48, 20.48]
365.53999999999996
18.276999999999997
{'CPU': 'Altra', 'CORES': 1, 'ITERATIONS': 100, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 43.16092610359192, 'TIME_S_1KI': 431.6092610359192, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1189.33110124588, 'W': 22.848960166729626, 'J_1KI': 11893.3110124588, 'W_1KI': 228.48960166729626, 'W_D': 4.571960166729628, 'J_D': 237.97907564592364, 'W_D_1KI': 45.71960166729628, 'J_D_1KI': 457.19601667296286}

1
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 1, "ITERATIONS": 362, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.810595989227295, "TIME_S_1KI": 29.863524832119598, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1183.753153848648, "W": 76.66, "J_1KI": 3270.036336598475, "W_1KI": 211.76795580110496, "W_D": 41.3065, "J_D": 637.8385031235218, "W_D_1KI": 114.10635359116023, "J_D_1KI": 315.21092152254204}

74
pytorch/output_389000+_1core/epyc_7313p_1_csr_10_10_10_msdoor.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', '100', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 2.8945302963256836}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0122, 0.2882, 0.1159, ..., 0.4750, 0.0340, 0.6603])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 2.8945302963256836 seconds
['apptainer', 'run', '--env', 'OMP_PROC_BIND=true', '--env', 'OMP_PLACES={0:1}', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '362', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.810595989227295}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.2571, 0.3344, 0.1194, ..., 0.8187, 0.7037, 0.4726])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.810595989227295 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.2571, 0.3344, 0.1194, ..., 0.8187, 0.7037, 0.4726])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.810595989227295 seconds
[41.31, 42.41, 38.93, 38.76, 38.75, 38.61, 39.07, 40.32, 39.22, 39.05]
[76.66]
15.441601276397705
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 362, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.810595989227295, 'TIME_S_1KI': 29.863524832119598, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1183.753153848648, 'W': 76.66}
[41.31, 42.41, 38.93, 38.76, 38.75, 38.61, 39.07, 40.32, 39.22, 39.05, 39.97, 39.08, 38.84, 38.82, 38.78, 38.68, 38.9, 38.82, 39.3, 39.23]
707.0699999999999
35.3535
{'CPU': 'Epyc 7313P', 'CORES': 1, 'ITERATIONS': 362, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.810595989227295, 'TIME_S_1KI': 29.863524832119598, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1183.753153848648, 'W': 76.66, 'J_1KI': 3270.036336598475, 'W_1KI': 211.76795580110496, 'W_D': 41.3065, 'J_D': 637.8385031235218, 'W_D_1KI': 114.10635359116023, 'J_D_1KI': 315.21092152254204}

1
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 1, "ITERATIONS": 181, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.025615215301514, "TIME_S_1KI": 55.390139311058086, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1684.3236584544181, "W": 50.05, "J_1KI": 9305.655571571371, "W_1KI": 276.5193370165745, "W_D": 32.91325, "J_D": 1107.6236893431544, "W_D_1KI": 181.84116022099448, "J_D_1KI": 1004.6472940386435}

74
pytorch/output_389000+_1core/xeon_4216_1_csr_10_10_10_msdoor.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', '100', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 5.796452283859253}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.5016, 0.6412, 0.4729, ..., 0.2262, 0.2092, 0.6154])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 5.796452283859253 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', '181', '-m', 'matrices/389000+_cols/msdoor.mtx', '-c', '1']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.025615215301514}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.6635, 0.3125, 0.6322, ..., 0.2814, 0.2894, 0.9983])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.025615215301514 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.6635, 0.3125, 0.6322, ..., 0.2814, 0.2894, 0.9983])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.025615215301514 seconds
[20.79, 18.75, 18.84, 18.78, 18.62, 18.79, 22.54, 18.49, 18.38, 20.09]
[50.05]
33.652820348739624
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 181, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.025615215301514, 'TIME_S_1KI': 55.390139311058086, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1684.3236584544181, 'W': 50.05}
[20.79, 18.75, 18.84, 18.78, 18.62, 18.79, 22.54, 18.49, 18.38, 20.09, 18.86, 18.34, 18.49, 18.63, 18.54, 19.75, 19.38, 18.65, 18.68, 18.43]
342.735
17.13675
{'CPU': 'Xeon 4216', 'CORES': 1, 'ITERATIONS': 181, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.025615215301514, 'TIME_S_1KI': 55.390139311058086, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1684.3236584544181, 'W': 50.05, 'J_1KI': 9305.655571571371, 'W_1KI': 276.5193370165745, 'W_D': 32.91325, 'J_D': 1107.6236893431544, 'W_D_1KI': 181.84116022099448, 'J_D_1KI': 1004.6472940386435}

1
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Altra", "CORES": 80, "ITERATIONS": 100, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 11.470221519470215, "TIME_S_1KI": 114.70221519470215, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 957.8876025772095, "W": 54.677060171971085, "J_1KI": 9578.876025772095, "W_1KI": 546.7706017197108, "W_D": 36.076060171971086, "J_D": 632.016620496273, "W_D_1KI": 360.76060171971085, "J_D_1KI": 3607.6060171971085}

51
pytorch/output_389000+_maxcore/altra_max_csr_10_10_10_msdoor.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 100 -m matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 11.470221519470215}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0348, 0.4239, 0.1578, ..., 0.9250, 0.1408, 0.2700])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 11.470221519470215 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0348, 0.4239, 0.1578, ..., 0.9250, 0.1408, 0.2700])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 11.470221519470215 seconds
[20.56, 20.56, 20.52, 20.48, 20.68, 20.68, 20.56, 20.6, 20.64, 20.56]
[20.56, 20.64, 21.64, 22.44, 26.08, 26.08, 35.0, 38.96, 51.72, 68.64, 76.08, 87.88, 94.36, 93.12, 94.56, 93.48, 95.12]
17.51900339126587
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 100, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 11.470221519470215, 'TIME_S_1KI': 114.70221519470215, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 957.8876025772095, 'W': 54.677060171971085}
[20.56, 20.56, 20.52, 20.48, 20.68, 20.68, 20.56, 20.6, 20.64, 20.56, 20.84, 20.56, 20.68, 20.64, 20.6, 20.6, 20.88, 20.96, 20.96, 20.88]
372.02
18.601
{'CPU': 'Altra', 'CORES': 80, 'ITERATIONS': 100, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 11.470221519470215, 'TIME_S_1KI': 114.70221519470215, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 957.8876025772095, 'W': 54.677060171971085, 'J_1KI': 9578.876025772095, 'W_1KI': 546.7706017197108, 'W_D': 36.076060171971086, 'J_D': 632.016620496273, 'W_D_1KI': 360.76060171971085, 'J_D_1KI': 3607.6060171971085}

1
pytorch/output_389000+_maxcore/epyc_7313p_max_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Epyc 7313P", "CORES": 16, "ITERATIONS": 2238, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 11.481182098388672, "TIME_S_1KI": 5.130108176223714, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 1893.635374045372, "W": 130.54, "J_1KI": 846.1284066333208, "W_1KI": 58.32886505808758, "W_D": 94.69824999999999, "J_D": 1373.7088713052867, "W_D_1KI": 42.313784629133146, "J_D_1KI": 18.90696364125699}

120
pytorch/output_389000+_maxcore/epyc_7313p_max_csr_10_10_10_msdoor.output Normal file
View File

@ -0,0 +1,120 @@
['apptainer', 'run', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 0.544938325881958}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.0497, 0.3572, 0.9272, ..., 0.8625, 0.2792, 0.5285])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 0.544938325881958 seconds
['apptainer', 'run', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '1926', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 9.506705045700073}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.9110, 0.2003, 0.2096, ..., 0.2036, 0.8917, 0.5683])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 9.506705045700073 seconds
['apptainer', 'run', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2127', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 9.97889494895935}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.5968, 0.7420, 0.6758, ..., 0.0056, 0.9599, 0.3333])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 9.97889494895935 seconds
['apptainer', 'run', 'pytorch-epyc_7313p.sif', 'python3', 'spmv.py', 'suitesparse', 'csr', '2238', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 11.481182098388672}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.1118, 0.5594, 0.9085, ..., 0.4616, 0.1638, 0.5506])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 11.481182098388672 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.1118, 0.5594, 0.9085, ..., 0.4616, 0.1638, 0.5506])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 11.481182098388672 seconds
[40.88, 39.48, 39.55, 40.68, 39.35, 39.35, 39.89, 39.72, 39.69, 39.48]
[130.54]
14.506169557571411
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2238, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 11.481182098388672, 'TIME_S_1KI': 5.130108176223714, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1893.635374045372, 'W': 130.54}
[40.88, 39.48, 39.55, 40.68, 39.35, 39.35, 39.89, 39.72, 39.69, 39.48, 40.3, 39.34, 40.16, 39.39, 39.51, 39.33, 39.34, 39.93, 39.57, 44.45]
716.835
35.841750000000005
{'CPU': 'Epyc 7313P', 'CORES': 16, 'ITERATIONS': 2238, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 11.481182098388672, 'TIME_S_1KI': 5.130108176223714, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 1893.635374045372, 'W': 130.54, 'J_1KI': 846.1284066333208, 'W_1KI': 58.32886505808758, 'W_D': 94.69824999999999, 'J_D': 1373.7088713052867, 'W_D_1KI': 42.313784629133146, 'J_D_1KI': 18.90696364125699}

1
pytorch/output_389000+_maxcore/xeon_4216_max_csr_10_10_10_msdoor.json Normal file
View File

@ -0,0 +1 @@
{"CPU": "Xeon 4216", "CORES": 16, "ITERATIONS": 1379, "MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.283698081970215, "TIME_S_1KI": 7.457359015206827, "BASELINE_TIME_S": 10, "BASELINE_DELAY_S": 10, "J": 2003.8670825815202, "W": 66.59, "J_1KI": 1453.1305892541843, "W_1KI": 48.288614938361135, "W_D": 49.53425, "J_D": 1490.615002783656, "W_D_1KI": 35.920413343002174, "J_D_1KI": 26.04816050979128}

97
pytorch/output_389000+_maxcore/xeon_4216_max_csr_10_10_10_msdoor.output Normal file
View File

@ -0,0 +1,97 @@
['apptainer', 'run', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '100', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 1.2347450256347656}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.4403, 0.7085, 0.5973, ..., 0.5516, 0.0584, 0.9044])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 1.2347450256347656 seconds
['apptainer', 'run', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '850', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 6.469892501831055}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.6729, 0.8216, 0.8655, ..., 0.6622, 0.9214, 0.0755])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 6.469892501831055 seconds
['apptainer', 'run', 'pytorch-xeon_4216.sif', 'numactl', '--cpunodebind=0', '--membind=0', 'python3', 'spmv.py', 'suitesparse', 'csr', '1379', '-m', 'matrices/389000+_cols/msdoor.mtx']
{"MATRIX_TYPE": "SuiteSparse", "MATRIX_FILE": "msdoor", "MATRIX_FORMAT": "csr", "MATRIX_SHAPE": [415863, 415863], "MATRIX_ROWS": 415863, "MATRIX_SIZE": 172942034769, "MATRIX_NNZ": 20240935, "MATRIX_DENSITY": 0.00011703883921012015, "TIME_S": 10.283698081970215}
/nfshomes/vut/ampere_research/pytorch/spmv.py:75: UserWarning: Sparse CSR tensor support is in beta state. If 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.9393, 0.6181, 0.5648, ..., 0.9318, 0.7600, 0.4697])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.283698081970215 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, 32, 64, ..., 20240851,
20240893, 20240935]),
col_indices=tensor([ 0, 1, 2, ..., 415860, 415861,
415862]),
values=tensor([1732682.0000, 32540.3633, 24619.0176, ...,
-97620.4609, -360329.0312, 2075205.5000]),
size=(415863, 415863), nnz=20240935, layout=torch.sparse_csr)
tensor([0.9393, 0.6181, 0.5648, ..., 0.9318, 0.7600, 0.4697])
Matrix Type: SuiteSparse
Matrix: msdoor
Matrix Format: csr
Shape: torch.Size([415863, 415863])
Rows: 415863
Size: 172942034769
NNZ: 20240935
Density: 0.00011703883921012015
Time: 10.283698081970215 seconds
[19.74, 19.62, 18.69, 18.65, 18.95, 18.9, 18.69, 18.4, 18.92, 18.85]
[66.59]
30.092612743377686
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1379, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.283698081970215, 'TIME_S_1KI': 7.457359015206827, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2003.8670825815202, 'W': 66.59}
[19.74, 19.62, 18.69, 18.65, 18.95, 18.9, 18.69, 18.4, 18.92, 18.85, 19.77, 18.65, 19.56, 18.53, 18.44, 18.48, 18.72, 20.39, 19.13, 18.43]
341.115
17.05575
{'CPU': 'Xeon 4216', 'CORES': 16, 'ITERATIONS': 1379, 'MATRIX_TYPE': 'SuiteSparse', 'MATRIX_FILE': 'msdoor', 'MATRIX_FORMAT': 'csr', 'MATRIX_SHAPE': [415863, 415863], 'MATRIX_ROWS': 415863, 'MATRIX_SIZE': 172942034769, 'MATRIX_NNZ': 20240935, 'MATRIX_DENSITY': 0.00011703883921012015, 'TIME_S': 10.283698081970215, 'TIME_S_1KI': 7.457359015206827, 'BASELINE_TIME_S': 10, 'BASELINE_DELAY_S': 10, 'J': 2003.8670825815202, 'W': 66.59, 'J_1KI': 1453.1305892541843, 'W_1KI': 48.288614938361135, 'W_D': 49.53425, 'J_D': 1490.615002783656, 'W_D_1KI': 35.920413343002174, 'J_D_1KI': 26.04816050979128}