cephi/ampere_research
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1735ecf142
ampere_research/pytorch/output/altra_10_30_sx-mathoverflow_1000.output
cephi 93690abfee new output
2024-12-03 08:53:39 -05:00

164 lines
9.8 KiB
Plaintext
Raw Blame History

srun: Job time limit was unset; set to partition default of 60 minutes
srun: ################################################################################
srun: # Please note that the oasis compute nodes have aarch64 architecture CPUs. #
srun: # All submission nodes and all other compute nodes have x86_64 architecture #
srun: # CPUs. Programs, environments, or other software that was built on x86_64 #
srun: # nodes may need to be rebuilt to properly execute on these nodes. #
srun: ################################################################################
srun: job 3394987 queued and waiting for resources
srun: job 3394987 has been allocated resources
/nfshomes/vut/ampere_research/pytorch/spmv.py:20: UserWarning: Sparse CSR tensor support is in beta state. If 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.)
).to_sparse_csr().type(torch.float)
tensor(crow_indices=tensor([ 0, 317, 416, ..., 239976, 239977,
239978]),
col_indices=tensor([ 0, 1, 2, ..., 1483, 2179, 24817]),
values=tensor([151., 17., 6., ..., 1., 1., 1.]),
size=(24818, 24818), nnz=239978, layout=torch.sparse_csr)
tensor([0.8864, 0.5637, 0.9805, ..., 0.0234, 0.9487, 0.4860])
Matrix: sx-mathoverflow
Shape: torch.Size([24818, 24818])
NNZ: 239978
Density: 0.00038961697406616504
Time: 5.484489917755127 seconds
Performance counter stats for 'apptainer run pytorch-altra.sif -c numactl --cpunodebind=0 --membind=0 python spmv.py matrices/sx-mathoverflow.mtx 1000':
50.36 msec task-clock:u # 0.006 CPUs utilized
0 context-switches:u # 0.000 /sec
0 cpu-migrations:u # 0.000 /sec
3,296 page-faults:u # 65.452 K/sec
56,049,457 cycles:u # 1.113 GHz (49.66%)
72,333,565 instructions:u # 1.29 insn per cycle (66.35%)
<not supported> branches:u
369,218 branch-misses:u (86.12%)
33,730,437 L1-dcache-loads:u # 669.814 M/sec (93.88%)
459,922 L1-dcache-load-misses:u # 1.36% of all L1-dcache accesses
<not supported> LLC-loads:u
<not supported> LLC-load-misses:u
31,827,672 L1-icache-loads:u # 632.030 M/sec
295,060 L1-icache-load-misses:u # 0.93% of all L1-icache accesses
54,366,618 dTLB-loads:u # 1.080 G/sec (35.64%)
84,768 dTLB-load-misses:u # 0.16% of all dTLB cache accesses (25.48%)
12,107,953 iTLB-loads:u # 240.438 M/sec (10.11%)
<not counted> iTLB-load-misses:u (0.00%)
8.968532171 seconds time elapsed
20.749643000 seconds user
28.745486000 seconds sys
/nfshomes/vut/ampere_research/pytorch/spmv.py:20: UserWarning: Sparse CSR tensor support is in beta state. If 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.)
).to_sparse_csr().type(torch.float)
tensor(crow_indices=tensor([ 0, 317, 416, ..., 239976, 239977,
239978]),
col_indices=tensor([ 0, 1, 2, ..., 1483, 2179, 24817]),
values=tensor([151., 17., 6., ..., 1., 1., 1.]),
size=(24818, 24818), nnz=239978, layout=torch.sparse_csr)
tensor([0.5549, 0.0336, 0.9472, ..., 0.2657, 0.3394, 0.6185])
Matrix: sx-mathoverflow
Shape: torch.Size([24818, 24818])
NNZ: 239978
Density: 0.00038961697406616504
Time: 5.532417297363281 seconds
Performance counter stats for 'apptainer run pytorch-altra.sif -c numactl --cpunodebind=0 --membind=0 python spmv.py matrices/sx-mathoverflow.mtx 1000':
325,529 BR_MIS_PRED_RETIRED:u # 0.0 per branch branch_misprediction_ratio
19,463,406 BR_RETIRED:u
8.912497962 seconds time elapsed
20.214519000 seconds user
31.566513000 seconds sys
/nfshomes/vut/ampere_research/pytorch/spmv.py:20: UserWarning: Sparse CSR tensor support is in beta state. If 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.)
).to_sparse_csr().type(torch.float)
tensor(crow_indices=tensor([ 0, 317, 416, ..., 239976, 239977,
239978]),
col_indices=tensor([ 0, 1, 2, ..., 1483, 2179, 24817]),
values=tensor([151., 17., 6., ..., 1., 1., 1.]),
size=(24818, 24818), nnz=239978, layout=torch.sparse_csr)
tensor([0.3330, 0.8843, 0.5150, ..., 0.7292, 0.0873, 0.4184])
Matrix: sx-mathoverflow
Shape: torch.Size([24818, 24818])
NNZ: 239978
Density: 0.00038961697406616504
Time: 5.457342863082886 seconds
Performance counter stats for 'apptainer run pytorch-altra.sif -c numactl --cpunodebind=0 --membind=0 python spmv.py matrices/sx-mathoverflow.mtx 1000':
27,374,917 L1I_TLB:u # 0.0 per TLB access itlb_walk_ratio
5,203 ITLB_WALK:u
16,771 DTLB_WALK:u # 0.0 per TLB access dtlb_walk_ratio
36,373,182 L1D_TLB:u
8.730534933 seconds time elapsed
20.156482000 seconds user
31.426118000 seconds sys
/nfshomes/vut/ampere_research/pytorch/spmv.py:20: UserWarning: Sparse CSR tensor support is in beta state. If 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.)
).to_sparse_csr().type(torch.float)
tensor(crow_indices=tensor([ 0, 317, 416, ..., 239976, 239977,
239978]),
col_indices=tensor([ 0, 1, 2, ..., 1483, 2179, 24817]),
values=tensor([151., 17., 6., ..., 1., 1., 1.]),
size=(24818, 24818), nnz=239978, layout=torch.sparse_csr)
tensor([0.5864, 0.4449, 0.4042, ..., 0.1651, 0.7793, 0.8302])
Matrix: sx-mathoverflow
Shape: torch.Size([24818, 24818])
NNZ: 239978
Density: 0.00038961697406616504
Time: 5.449937582015991 seconds
Performance counter stats for 'apptainer run pytorch-altra.sif -c numactl --cpunodebind=0 --membind=0 python spmv.py matrices/sx-mathoverflow.mtx 1000':
31,839,975 L1I_CACHE:u # 0.0 per cache access l1i_cache_miss_ratio
274,158 L1I_CACHE_REFILL:u
471,992 L1D_CACHE_REFILL:u # 0.0 per cache access l1d_cache_miss_ratio
33,638,817 L1D_CACHE:u
8.845491835 seconds time elapsed
20.577696000 seconds user
35.105662000 seconds sys
/nfshomes/vut/ampere_research/pytorch/spmv.py:20: UserWarning: Sparse CSR tensor support is in beta state. If 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.)
).to_sparse_csr().type(torch.float)
tensor(crow_indices=tensor([ 0, 317, 416, ..., 239976, 239977,
239978]),
col_indices=tensor([ 0, 1, 2, ..., 1483, 2179, 24817]),
values=tensor([151., 17., 6., ..., 1., 1., 1.]),
size=(24818, 24818), nnz=239978, layout=torch.sparse_csr)
tensor([0.8880, 0.4700, 0.5542, ..., 0.8505, 0.9123, 0.5742])
Matrix: sx-mathoverflow
Shape: torch.Size([24818, 24818])
NNZ: 239978
Density: 0.00038961697406616504
Time: 5.400304794311523 seconds
Performance counter stats for 'apptainer run pytorch-altra.sif -c numactl --cpunodebind=0 --membind=0 python spmv.py matrices/sx-mathoverflow.mtx 1000':
538,067 LL_CACHE_MISS_RD:u # 1.0 per cache access ll_cache_read_miss_ratio
557,981 LL_CACHE_RD:u
170,169 L2D_TLB:u # 0.1 per TLB access l2_tlb_miss_ratio
21,987 L2D_TLB_REFILL:u
301,746 L2D_CACHE_REFILL:u # 0.2 per cache access l2_cache_miss_ratio
1,735,872 L2D_CACHE:u
8.606800178 seconds time elapsed
21.064990000 seconds user
34.158762000 seconds sys
Reference in New Issue View Git Blame Copy Permalink