Skip to content
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

ci: Switch to GitHub-hosted GPU runner #283

Merged
merged 2 commits into from
Dec 4, 2024
Merged

ci: Switch to GitHub-hosted GPU runner #283

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e48a8da
ci: Switch to GitHub-hosted GPU runner
samuelburnham Dec 4, 2024
518aa16
Prep for review
samuelburnham Dec 4, 2024
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
54 changes: 34 additions & 20 deletions .github/workflows/rust.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,13 +80,11 @@ jobs:
licenses-audits:
uses: argumentcomputer/ci-workflows/.github/workflows/licenses-audits.yml@main

# Runs the test suite on a self-hosted GPU machine with CUDA enabled
# Runs the test suite on a GPU machine with CUDA enabled
test-cuda:
name: Rust tests on CUDA
runs-on: self-hosted
runs-on: gpu-ci-t4
env:
NVIDIA_VISIBLE_DEVICES: all
NVIDIA_DRIVER_CAPABILITITES: compute,utility
EC_GPU_FRAMEWORK: cuda
steps:
- uses: actions/checkout@v4
Expand All @@ -99,6 +97,20 @@ jobs:
- uses: Swatinem/rust-cache@v2
# Check we have access to the machine's Nvidia drivers
- run: nvidia-smi
- name: Install CUDA
run: |
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-4
echo "PATH=/usr/local/cuda/bin:$PATH" >> $GITHUB_ENV
# Check that CUDA is installed with a driver-compatible version
# This must also be compatible with the GPU architecture, see below comment
- run: nvcc --version
- name: Install deps
run: |
sudo apt-get update
sudo apt-get install -y build-essential
# The `compute`/`sm` number corresponds to the Nvidia GPU architecture
# In this case, the self-hosted machine uses the Ampere architecture, but we want this to be configurable
# See https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/
Expand All @@ -107,20 +119,15 @@ jobs:
- name: set env for EC_GPU
run: echo 'EC_GPU_CUDA_NVCC_ARGS=--fatbin --gpu-architecture=sm_${{ env.CUDA_ARCH }} --generate-code=arch=compute_${{ env.CUDA_ARCH }},code=sm_${{ env.CUDA_ARCH }}' >> $GITHUB_ENV
- run: echo "${{ env.EC_GPU_CUDA_NVCC_ARGS}}"
# Check that CUDA is installed with a driver-compatible version
# This must also be compatible with the GPU architecture, see above link
- run: nvcc --version
- name: CUDA tests
run: |
cargo nextest run --release --no-default-features --features std,cuda,pasta,bls,arity2,arity4,arity8,arity11,arity16,arity24,arity36

# Runs the test suite on a self-hosted GPU machine with CUDA and OpenCL enabled (that is using the OpenCL backend for NVIDIA GPUs)
# Runs the test suite on a GPU machine with CUDA and OpenCL enabled (that is using the OpenCL backend for NVIDIA GPUs)
test-opencl:
name: Rust tests on OpenCL
runs-on: self-hosted
runs-on: gpu-ci-t4
env:
NVIDIA_VISIBLE_DEVICES: all
NVIDIA_DRIVER_CAPABILITITES: compute,utility
EC_GPU_FRAMEWORK: opencl
steps:
- uses: actions/checkout@v4
Expand All @@ -131,12 +138,24 @@ jobs:
- uses: dtolnay/rust-toolchain@stable
- uses: taiki-e/install-action@nextest
- uses: Swatinem/rust-cache@v2
- name: Install GPU deps
run: |
apt-get update
apt-get -y install ocl-icd-opencl-dev
# Check we have access to the machine's Nvidia drivers
- run: nvidia-smi
- name: Install CUDA
run: |
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-4
echo "PATH=/usr/local/cuda/bin:$PATH" >> $GITHUB_ENV
# Check that CUDA is installed with a driver-compatible version
# This must also be compatible with the GPU architecture, see below comments
- run: nvcc --version
- name: Install deps
run: |
sudo apt-get update
sudo apt-get -y install build-essential ocl-icd-opencl-dev clinfo
# Check that we can access the OpenCL headers
- run: clinfo
# The `compute`/`sm` number corresponds to the Nvidia GPU architecture
# In this case, the self-hosted machine uses the Ampere architecture, but we want this to be configurable
# See https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/
Expand All @@ -145,11 +164,6 @@ jobs:
- name: set env for EC_GPU
run: echo 'EC_GPU_CUDA_NVCC_ARGS=--fatbin --gpu-architecture=sm_${{ env.CUDA_ARCH }} --generate-code=arch=compute_${{ env.CUDA_ARCH }},code=sm_${{ env.CUDA_ARCH }}' >> $GITHUB_ENV
- run: echo "${{ env.EC_GPU_CUDA_NVCC_ARGS}}"
# Check that CUDA is installed with a driver-compatible version
# This must also be compatible with the GPU architecture, see above link
- run: nvcc --version
# Check that we can access the OpenCL headers
- run: clinfo
- name: OpenCL tests
run: |
cargo nextest run --release --no-default-features --features std,strengthened,abomonation,opencl,pasta,bls,arity2,arity4,arity8,arity11,arity16,arity24,arity36
Loading