biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix][Build/CI] Fix sparse CUTLASS compilation on CUDA [12.0, 12.2) (#11311)

Browse Source 
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
This commit is contained in:
Tyler Michael Smith
2024-12-18 21:43:30 -05:00
committed by GitHub
parent fdea8ec167
commit 5a9da2e6e9
12 changed files with 89 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
csrc/torch_bindings.cpp
View File

@@ -321,6 +321,13 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
ops.def("cutlass_scaled_mm_supports_fp8(int cuda_device_capability) -> bool");
ops.impl("cutlass_scaled_mm_supports_fp8", &cutlass_scaled_mm_supports_fp8);
// Check if cutlass sparse scaled_mm is supported for CUDA devices of the
// given capability
ops.def(
"cutlass_sparse_scaled_mm_supported(int cuda_device_capability) -> bool");
ops.impl("cutlass_sparse_scaled_mm_supported",
&cutlass_sparse_scaled_mm_supported);
// CUTLASS sparse GEMM, supporting symmetric per-tensor or per-row/column
// quantization, as well as bias
ops.def(