chenchenghao/DeepGEMM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Multiple updates and refactorings (#231)

Browse Source
This commit is contained in:
Ray Wang
2025-11-21 17:49:47 +08:00
committed by GitHub
parent bb4424aad4
commit 38f8ef73a4
80 changed files with 3767 additions and 2103 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+2 -4
View File
@@ -27,7 +27,7 @@ Despite its lightweight design, DeepGEMM's performance matches or exceeds expert
- [x] Fix TMA multicast compatibility for indivisible shapes
- [x] Skip useless computation on M
- [x] NVRTC as a faster compiler
- [ ] Sanitizer for testing
- [x] Sanitizer for testing
- [x] Weight gradient kernels for dense models
- [x] Weight gradient kernels for MoE models
- [ ] Better `get_best_configs` modeling
@@ -69,9 +69,7 @@ cat develop.sh
# Test all GEMM implements
python tests/test_layout.py
python tests/test_attention.py
python tests/test_bf16.py
python tests/test_fp8.py
python tests/test_lazy_init.py
python tests/test_core.py
```
### Installation