pyc-kernels / PERFORMANCE_SUMMARY.md
Allanatrix's picture
Publish PyC CUDA kernel lab and performance summary
5c49704 verified
|
Raw
History Blame Contribute Delete
826 Bytes
# Performance summary
These figures are selected from the PyC H100 kernel-lab campaign and are
included as orientation, not as universal benchmark claims.
| Lane | Shape | Result |
| --- | --- | --- |
| Hopper WMMA FP16 | 1024^3 | 0.105 ms, 20.473 TFLOPS |
| Hopper WMMA BF16 | 1024^3 | 0.102 ms, 21.024 TFLOPS |
| Hopper cuBLASLt BF16 control | 4096^3 | 0.162 ms, 846.466 TFLOPS |
| Hopper async square K64 | 4096^3 | 0.9252 ms in the captured profile |
The engineering progression is: shared-memory tiling and reuse, WMMA Tensor
Core execution, BF16/FP16 comparison, `cp.async` double buffering, CTA shape
and warp-work assignment, K-stage depth, and finally a cuBLASLt control lane.
Always interpret a result together with GPU model, architecture, CUDA version,
matrix shape, warmup/repeat policy, and correctness mode.