Skip to content

Adds infra to use NCCL from pypi #1184

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

Closed
wants to merge 1 commit into from
Closed

Adds infra to use NCCL from pypi #1184

Show file tree
Hide file tree
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
12 changes: 11 additions & 1 deletion common/install_cuda.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,7 +98,7 @@ function install_116 {
}

function install_117 {
echo "Installing CUDA 11.7 and CuDNN 8.3"
echo "Installing CUDA 11.7 and CuDNN 8.3 and NCCL 2.14"
rm -rf /usr/local/cuda-11.7 /usr/local/cuda
# install CUDA 11.7.0 in the same container
wget -q https://developer.download.nvidia.com/compute/cuda/11.7.0/local_installers/cuda_11.7.0_515.43.04_linux.run
Expand All @@ -116,6 +116,16 @@ function install_117 {
cd ..
rm -rf tmp_cudnn
ldconfig

# NCCL license: https://docs.nvidia.com/deeplearning/nccl/#licenses
mkdir tmp_nccl && cd tmp_nccl
wget -q https://developer.download.nvidia.com/compute/redist/nccl/v2.14/nccl_2.14.3-1+cuda11.7_x86_64.txz
tar xf nccl_2.14.3-1+cuda11.7_x86_64.txz
cp -a nccl_2.14.3-1+cuda11.7_x86_64/include/* /usr/local/cuda/include/
cp -a nccl_2.14.3-1+cuda11.7_x86_64/lib/* /usr/local/cuda/lib64/
cd ..
rm -rf tmp_nccl
ldconfig
}

function prune_102 {
Expand Down
5 changes: 4 additions & 1 deletion manywheel/build_cuda.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -275,15 +275,18 @@ elif [[ $CUDA_VERSION == "11.7" ]]; then
"libcublasLt.so.11"
)
else
echo "Using cudnn and cublas from pypi."
echo "Using cudnn, cublas, and nccl from pypi."
CUDA_RPATHS=(
'$ORIGIN/../../nvidia/cublas/lib'
'$ORIGIN/../../nvidia/cudnn/lib'
'$ORIGIN/../../nvidia/nccl/lib'
)
CUDA_RPATHS=$(IFS=: ; echo "${CUDA_RPATHS[*]}")
export C_SO_RPATH=$CUDA_RPATHS':$ORIGIN:$ORIGIN/lib'
export LIB_SO_RPATH=$CUDA_RPATHS':$ORIGIN'
export FORCE_RPATH="--force-rpath"
export USE_STATIC_NCCL=0
export USE_SYSTEM_NCCL=1
fi
else
echo "Unknown cuda version $CUDA_VERSION"
Expand Down