Chenggang Zhao
55ab91f72f
Update performance
2025-03-25 18:06:47 +08:00
Chenggang Zhao
09d097f84d
Add some notes
2025-03-25 17:41:49 +08:00
Chenggang Zhao
25db8de345
Better performance
2025-03-25 17:34:06 +08:00
Chenggang Zhao
1999d553e5
Lower TMA requirement
2025-03-25 17:18:53 +08:00
Chenggang Zhao
ddccb230ca
Fix NVCC branch divergence
2025-03-25 17:12:51 +08:00
Chenggang Zhao
9c4f6f53f5
Optimize compilation speed
2025-03-25 16:51:21 +08:00
Chenggang Zhao
612dd57001
Simplify code
2025-03-25 16:45:20 +08:00
Chenggang Zhao
046fab64b7
Fix grouped GEMM cases
2025-03-25 16:41:44 +08:00
Chenggang Zhao
7768319ffe
Remove unaligned predicates
2025-03-25 16:32:40 +08:00
Chenggang Zhao
3497428a5e
Minor fix
2025-03-25 15:16:26 +08:00
Chenggang Zhao
7ffb118e54
Support multicasting on B
2025-03-25 14:56:42 +08:00
Chenggang Zhao
742fb1c8a5
Compilation-time GCD
2025-03-25 13:41:28 +08:00
Chenggang Zhao
b922e64cb2
Support block size 160
2025-03-25 13:37:59 +08:00
sazc
46eb0d08fb
Performance: Larger BlockTile optimizations enable 1470+ TFLOPS FP8 performance on the H800-SXM platform
2025-03-25 10:44:57 +08:00
Liang
3b3783d06c
Merge pull request #68 from ademeure/flush_l2_pr
...
Correctly flush L2 (+performance impact & upcoming optimization fork)
2025-03-16 09:16:34 +08:00
ademeure
6cbff5778f
Correctly flush L2, as reconstructing the tensors on every iteration effectively put them in the L2, and gave the GPU enough idle time to avoid thermal throttling in a potentially unrealistic way.
...
The previous behaviour is potentially representative of some use cases (e.g. previous kernel filling L2 with the data in a very specific way) but not standard benchmarking practice.
2025-03-15 20:46:24 +00:00
Liang
e1c070fbef
Merge pull request #65 from Z-NAVY/main
...
Fix get_col_major_tma_aligned_tensor to handle 2-dimensional inputs
2025-03-14 13:50:08 +08:00
Liang
4377c4dc57
Merge pull request #63 from fzyzcjy/patch-2
...
Super tiny fix typo
2025-03-14 10:27:48 +08:00
z-navy
3f92607b98
Fix get_col_major_tma_aligned_tensor to handle 2-dimensional inputs
2025-03-13 22:15:16 +08:00
fzyzcjy
e7fff7ef0a
Update m_grouped_gemm.py
2025-03-13 22:09:15 +08:00
Chenggang Zhao
bd2a775528
Code format
2025-03-11 13:26:10 +08:00
Chenggang Zhao
5233bad1e9
Merge pull request #55 from sleepcoo/fix-cudagraph
...
fix cuda_graph rng check error
2025-03-11 13:25:35 +08:00
sleepcoo
723a00338e
fix cuda_graph rng check error
2025-03-11 12:40:42 +08:00
Chenggang Zhao
5e4badc577
Fix type lint
2025-03-10 13:10:16 +08:00
Chenggang Zhao
ba1e93a5c7
Merge pull request #44 from sazczmh/main
...
Performance: Configuration algorithms tuned to minimize the impact of tail effects, now up to 1402 TFLOPS
2025-03-10 13:08:03 +08:00
sazc
bed67b234c
Minor fix
2025-03-10 13:02:02 +08:00
sazc
ed278eddd3
formats: Optimize get_best_configs implementation
2025-03-10 12:56:14 +08:00
sazc
50cf26cc7c
Performance: Configuration algorithms tuned to minimize the impact of tail effects, now up to 1402 TFLOPS
2025-03-10 11:45:05 +08:00
Chenggang Zhao
39c10e6c31
Revert "Merge pull request #49 from A-transformer/maximum_fp8_e4m3_value"
...
This reverts commit 4d4f2342fe9d3222a93e
2025-03-10 09:47:02 +08:00
Liang
4d4f2342fe
Merge pull request #49 from A-transformer/maximum_fp8_e4m3_value
...
Maximum representable value in FP8 E4M3 format
2025-03-08 09:50:06 +08:00
A-transformer
629857685e
Maximum representable value in FP8 E4M3 format
...
Replace Hardcoded 448.0 with Global Constant FP8_E4M3_MAX for FP8 E4M3 Format
2025-03-07 19:58:02 +04:00
Liang
9d3222a93e
Merge pull request #42 from sazczmh/main
...
Performance: reducing the percentage of FFMA interleaving yields a sight performance gain, roughly 0.5%
2025-03-05 20:09:56 +08:00
sazc
fcd1dcd99d
Performance: reducing the percentage of FFMA interleaving yields a slight performance gain, roughly 0.5%
2025-03-05 17:50:22 +08:00
Chenggang Zhao
68fc742572
Merge pull request #36 from yz-tang/fix_setup_build
...
Fix setup build error when setuptools version is lower
2025-03-04 17:09:50 +08:00
yz-tang
6c59e0f40d
fix setup build error when setuptools version is lower
2025-03-04 16:53:00 +08:00
Chenggang Zhao
9b0dad8640
Add some notes for promotion
2025-03-04 11:42:20 +08:00
Liang
ded740f736
Fix documentation of m_grouped_gemm_fp8_fp8_bf16_nt_contiguous in m_grouped_gemm.py
2025-03-04 11:26:23 +08:00
Chenggang Zhao
dff6bb6f0b
Add some notes
2025-03-03 11:35:52 +08:00
Chenggang Zhao
6c5da03ba9
Support more shapes
2025-02-28 10:04:59 +08:00
Chenggang Zhao
b69f630b91
Minor fix util function
2025-02-28 09:46:38 +08:00
Chenggang Zhao
6e10cba207
Minor fix
2025-02-28 09:21:35 +08:00
Liang
fbec9e5eee
Update get_best_configs
...
a better strategy to choose config
2025-02-27 23:18:52 +08:00
Zhean Xu
461427ecd0
Merge pull request #27 from vatlor/main
...
fix typo
2025-02-27 20:37:31 +08:00
dotrail
488b5fc467
fix typo
2025-02-27 11:53:33 +00:00
Chenggang Zhao
6da94d2d36
Add extra TMA checks
2025-02-27 18:20:57 +08:00
Chenggang Zhao
ca13ce0fab
Fix TMA store bugs and code format
2025-02-27 17:57:21 +08:00
Chenggang Zhao
b05ed2f017
Code format
2025-02-27 10:50:20 +08:00
Chenggang Zhao
676329b8e2
Merge pull request #19 from dzhulgakov/fix-wheel
...
Fix wheel building
2025-02-27 10:44:11 +08:00
Chenggang Zhao
6e55da296f
Fix python -O mode issues
2025-02-27 10:42:46 +08:00
Chenggang Zhao
d5b974da2b
Merge pull request #16 from AcraeaTerpsicore/patch-1
...
Fix typos
2025-02-27 10:34:12 +08:00
