[Build] Fix DeepGEMM CUDA 12.9 release builds (#57554)
Signed-off-by: khluu <khluu000@gmail.com> Co-authored-by: OpenAI Codex <noreply@openai.com> Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
co-authored by
OpenAI Codex
Jee Jee Li
parent
86ce4d10e2
commit
eb87980585
@@ -29,10 +29,11 @@ if(DEEPGEMM_SRC_DIR)
|
||||
else()
|
||||
# Keep in sync with tools/install_deepgemm.sh
|
||||
set(_DEEPGEMM_UPSTREAM_REPO "https://github.com/vllm-project/DeepGEMM.git")
|
||||
# Pinned to the tip of the fork's dev branch: upstream 2.8.0 plus the SM120
|
||||
# Pinned to the fork's dev branch: upstream 2.8.0 plus the SM120
|
||||
# port, the SM90 paged-MQA kv_block=32/next_n=4 port, configurable SwiGLU
|
||||
# alpha/beta, and SiTU for FP8/FP4 Mega MoE.
|
||||
set(_DEEPGEMM_UPSTREAM_TAG "a6bbb8000161c0dc3a85a0300a905f76898a7913")
|
||||
# alpha/beta, and SiTU for FP8/FP4 Mega MoE. Also includes the CUDA 12.x
|
||||
# layout header fix from vllm-project/DeepGEMM#12.
|
||||
set(_DEEPGEMM_UPSTREAM_TAG "e1f418c2a4f20818221f6b0e578b4c2f634d4c3f")
|
||||
|
||||
set(_deepgemm_fc_root "${FETCHCONTENT_BASE_DIR}")
|
||||
if(NOT _deepgemm_fc_root)
|
||||
|
||||
@@ -7,10 +7,10 @@ set -e
|
||||
# Default values
|
||||
# Keep DEEPGEMM_GIT_REF in sync with cmake/external_projects/deepgemm.cmake
|
||||
DEEPGEMM_GIT_REPO="https://github.com/vllm-project/DeepGEMM.git"
|
||||
# NOTE: This targets the vLLM fork's dev branch tip, which carries the sm120
|
||||
# NOTE: This targets the vLLM fork's dev branch, which carries the sm120
|
||||
# and sm90 paged-MQA ports plus the SwiGLU alpha/beta and SiTU Mega MoE
|
||||
# activations
|
||||
DEEPGEMM_GIT_REF="a6bbb8000161c0dc3a85a0300a905f76898a7913"
|
||||
# activations, plus the CUDA 12.x layout header fix from vllm-project/DeepGEMM#12.
|
||||
DEEPGEMM_GIT_REF="e1f418c2a4f20818221f6b0e578b4c2f634d4c3f"
|
||||
WHEEL_DIR=""
|
||||
|
||||
# Parse command line arguments
|
||||
|
||||
Reference in New Issue
Block a user