Skip to content

TL/CUDA: Linear Broadcast for GPU #851

TL/CUDA: Linear Broadcast for GPU

TL/CUDA: Linear Broadcast for GPU #851

Workflow file for this run

name: HPC_SDK
on: [push, pull_request]
env:
HPCXDIR: /opt/nvidia/hpc_sdk/Linux_x86_64/2023/comm_libs/12.2/hpcx/latest/
NCCLDIR: /opt/nvidia/hpc_sdk/Linux_x86_64/2023/comm_libs/12.2/nccl/
CUDADIR: /opt/nvidia/hpc_sdk/Linux_x86_64/2023/cuda/12.2/
jobs:
build:
runs-on: ubuntu-20.04
container:
image: nvcr.io/nvidia/nvhpc:23.9-devel-cuda12.2-ubuntu22.04
steps:
- name: Install dependencies
run: |
apt-get update
apt-get install -y --no-install-recommends libiberty-dev
- uses: actions/checkout@v1
- name: Build UCC
run: |
./autogen.sh
CC=nvc CXX=nvc++ ./configure --with-tls=ucp,mlx5,cuda,self,nccl,sharp --with-mpi=${HPCXDIR}/ompi --with-sharp=${HPCXDIR}/sharp --with-ucx=${HPCXDIR}/ucx --with-cuda=${CUDADIR} --with-nccl=${NCCLDIR} --with-nvcc-gencode="-gencode=arch=compute_80,code=sm_80"
make -j`nproc` install