[Kernel] FlashMLA integration (#13747)

Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
This commit is contained in:
Lucas Wilkinson
2025-02-27 10:35:08 +08:00
committed by GitHub
parent b382a7f28f
commit f95903909f
11 changed files with 733 additions and 86 deletions
+6
View File
@@ -328,6 +328,7 @@ class repackage_wheel(build_ext):
files_to_copy = [
"vllm/_C.abi3.so",
"vllm/_moe_C.abi3.so",
"vllm/_flashmla_C.abi3.so",
"vllm/vllm_flash_attn/_vllm_fa2_C.abi3.so",
"vllm/vllm_flash_attn/_vllm_fa3_C.abi3.so",
"vllm/vllm_flash_attn/flash_attn_interface.py",
@@ -612,6 +613,11 @@ if _is_cuda():
# FA3 requires CUDA 12.0 or later
ext_modules.append(
CMakeExtension(name="vllm.vllm_flash_attn._vllm_fa3_C"))
if envs.VLLM_USE_PRECOMPILED or get_nvcc_cuda_version() >= Version("12.3"):
# Optional since this doesn't get built (produce an .so file) when
# not targeting a hopper system
ext_modules.append(
CMakeExtension(name="vllm._flashmla_C", optional=True))
ext_modules.append(CMakeExtension(name="vllm.cumem_allocator"))
if _build_custom_ops():