v4.0 update. (#2371)

This commit is contained in:
Junkai-Wu
2025-06-06 14:39:20 +08:00
committed by GitHub
parent 2e2af190bd
commit 8bdbfca682
254 changed files with 29751 additions and 1980 deletions

View File

@ -14,8 +14,8 @@ cmake $PATH -DCUTLASS_NVCC_ARCHS="100a" -DCUTLASS_ENABLE_GDC_FOR_SM100=1
### Minimum software
Like all other CUTLASS examples, the NVIDIA driver, runtime, and CUDA Toolkit are required.
This example specifically requires CUDA Toolkit 12.6 or newer, due to some of the necessary
CUDA graph APIs.
This example specifically requires CUDA Toolkit 12.8 or newer, since that is the first version
supporting the Blackwell architecture.
### Hardware / driver settings