Morph MORPH
SIGN IN SIGN UP
dhruvhead / DeepGEMM UNCLAIMED
0 0 2 Cuda

COMMITS

main
March 15, 2025
A
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.
ademeure committed
6cbff57
March 14, 2025
L
Merge pull request #65 from Z-NAVY/main
Liang committed
e1c070f
L
Merge pull request #63 from fzyzcjy/patch-2
Liang committed
4377c4d
March 13, 2025
Z
Fix get_col_major_tma_aligned_tensor to handle 2-dimensional inputs
z-navy committed
3f92607
F
Update m_grouped_gemm.py
fzyzcjy committed
e7fff7e
March 11, 2025
C
Code format
Chenggang Zhao committed
bd2a775
C
Merge pull request #55 from sleepcoo/fix-cudagraph
Chenggang Zhao committed
5233bad
S
fix cuda_graph rng check error
sleepcoo committed
723a003
March 10, 2025
C
Fix type lint
Chenggang Zhao committed
5e4badc
C
Merge pull request #44 from sazczmh/main
Chenggang Zhao committed
ba1e93a
S
Minor fix
sazc committed
bed67b2
S
formats: Optimize get_best_configs implementation
sazc committed
ed278ed
S
Performance: Configuration algorithms tuned to minimize the impact of tail effects, now up to 1402 TFLOPS
sazc committed
50cf26c
C
Revert "Merge pull request #49 from A-transformer/maximum_fp8_e4m3_value"
Chenggang Zhao committed
39c10e6
March 8, 2025
L
Merge pull request #49 from A-transformer/maximum_fp8_e4m3_value
Liang committed
4d4f234
March 7, 2025
A
Maximum representable value in FP8 E4M3 format
A-transformer committed
6298576
March 5, 2025
L
Merge pull request #42 from sazczmh/main
Liang committed
9d3222a
S
Performance: reducing the percentage of FFMA interleaving yields a slight performance gain, roughly 0.5%
sazc committed
fcd1dcd
March 4, 2025
C
Merge pull request #36 from yz-tang/fix_setup_build
Chenggang Zhao committed
68fc742
Y
fix setup build error when setuptools version is lower
yz-tang committed
6c59e0f
C
Add some notes for promotion
Chenggang Zhao committed
9b0dad8
L
Fix documentation of `m_grouped_gemm_fp8_fp8_bf16_nt_contiguous` in m_grouped_gemm.py
Liang committed
ded740f
March 3, 2025
C
Add some notes
Chenggang Zhao committed
dff6bb6
February 28, 2025
C
Support more shapes
Chenggang Zhao committed
6c5da03
C
Minor fix util function
Chenggang Zhao committed
b69f630
C
Minor fix
Chenggang Zhao committed
6e10cba
February 27, 2025
L
Update `get_best_configs`
Liang committed
fbec9e5
Z
Merge pull request #27 from vatlor/main
Zhean Xu committed
461427e
D
fix typo
dotrail committed
488b5fc
C
Add extra TMA checks
Chenggang Zhao committed
6da94d2