Skip to content

build: install flash_mla from source in the CI image #5481

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ko3n1g merged 1 commit into from
Jun 24, 2026
Merged

build: install flash_mla from source in the CI image #5481

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 10 additions & 1 deletion docker/Dockerfile.ci.dev
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,9 +40,18 @@ ENV NVTE_BUILD_NUM_PHILOX_ROUNDS=3
RUN --mount=type=cache,target=/root/.cache/uv \
bash -ex <<"EOF"
export NVTE_CUDA_ARCHS="80;90;100"
# flash-mla (no_pypi_wheels group) has no PyPI wheel and is built from source by uv. Point the
# compilers at the CCCL/libcu++ headers (under cccl/ in this base image, not the default CUDA
# include path) and scope the build to the target archs. Skipped for the LTS image.
FLASH_MLA_GROUP=""
if [ "$IMAGE_TYPE" != "lts" ]; then
FLASH_MLA_GROUP="--group no_pypi_wheels"
export FLASH_MLA_DISABLE_SM90=1 NVCC_THREADS=16 \
CFLAGS="-I/usr/local/cuda/include/cccl" CXXFLAGS="-I/usr/local/cuda/include/cccl"
fi
uv venv ${UV_PROJECT_ENVIRONMENT} --system-site-packages
uv sync --only-group build
uv sync --extra ${IMAGE_TYPE} --extra mlm --extra ssm --extra te --link-mode copy --locked \
uv sync --extra ${IMAGE_TYPE} --extra mlm --extra ssm --extra te ${FLASH_MLA_GROUP} --link-mode copy --locked \
--no-install-package torch \
--no-install-package torchvision \
--no-install-package triton \
Expand Down
Loading