Skip to content
KernelIndex
Search⌘K

PyTorch

python · BSD-3-Clause

Use it

Source available upstream · BSD-3-ClauseView source →

No install recipe recorded for this revision.

interfacepython
Compatibility
measured onNVIDIA H100
declared hardwaredeclared only
architectures—
dtypes

Benchmark evidence

12 measurements across 1 GPU, fastest first.

Operation / workload
Hardware
Latency
Rank
Observed
KL-divergence lossfp32 · [16384, 4096]
NVIDIA H100
1.33ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 8192]
NVIDIA H100
2.59ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 4096]
NVIDIA H100
2.86ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 16384]
NVIDIA H100
5.13ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 8192]
NVIDIA H100
5.59ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 32768]
NVIDIA H100
10.2ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 16384]
NVIDIA H100
11.1ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 65536]
NVIDIA H100
20.4ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 32768]
NVIDIA H100
23.1ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 131072]
NVIDIA H100
40.8ms
#2 of 2
2024-09-04
Show all 12 measurements ›
KL-divergence lossfp32 · [16384, 65536]
NVIDIA H100
46.1ms
#2 of 2
2024-09-04
KL-divergence lossfp32 · [16384, 131072]
NVIDIA H100
92.1ms
#2 of 2
2024-09-04

Reported · How evidence levels are derived →

Source and license

sourceavailable
revision digestsha256:9d1db6030b44918b1cda264dd2cd4e22a87e96c0b0e342ebd47437ce292a270c
license declaredBSD-3-Clause
license concludedBSD-3-Clause
imported2026-08-20

Best evidence level for this revision: reported

JSON