Skip to content

[CI] Make GitHub Build workflow fit the hosted runner (test PR) #49

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

Open
pggPL wants to merge 3 commits into
base: main
Choose a base branch
from
Open

[CI] Make GitHub Build workflow fit the hosted runner (test PR) #49

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1e3f521
[CI] Make GitHub Build workflow fit the hosted runner
pggPL Sep 8, 2026
81ba5e8
[CI] Build Core for every CUDA arch in a per-arch matrix; run on main…
pggPL Sep 8, 2026
a666ca5
[CI] Build PyTorch and JAX extensions for sm80 and sm90 to cover both…
pggPL Sep 9, 2026
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
68 changes: 46 additions & 22 deletions .github/workflows/build.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,15 +6,21 @@
name: 'Build'
on:
pull_request:
push:
branches: [main]
workflow_dispatch:
concurrency:
# Group by workflow name + PR number (for PRs) or ref (for branch/tag pushes)
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true
jobs:
core:
name: 'Core'
name: 'Core (sm${{ matrix.arch }})'
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
arch: ["70", "80", "89", "90", "100", "120"]
container:
image: nvcr.io/nvidia/cuda:12.9.2-devel-ubuntu22.04
options: --user root
Expand All @@ -37,14 +43,21 @@ jobs:
run: NVTE_USE_CCACHE=1 NVTE_CCACHE_BIN=sccache pip install --no-build-isolation . -v
env:
NVTE_FRAMEWORK: none
MAX_JOBS: 1
NVTE_CUDA_ARCHS: ${{ matrix.arch }}
MAX_JOBS: 2
SCCACHE_GHA_ENABLED: "true"
- name: 'Sanity check'
run: python3 -c "import transformer_engine"
run: |
ln -sf /usr/local/cuda/lib64/stubs/libcuda.so /usr/lib/x86_64-linux-gnu/libcuda.so.1
python3 -c "import transformer_engine"
working-directory: /
pytorch:
name: 'PyTorch'
name: 'PyTorch (sm${{ matrix.arch }})'
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
arch: ["80", "90"]
steps:
- name: Move /var/lib/docker/
shell: bash -euxo pipefail {0}
Expand Down Expand Up @@ -88,15 +101,22 @@ jobs:
'

- name: 'Build'
run: docker exec builder bash -c 'pip install --no-build-isolation . -v --no-deps'
env:
NVTE_FRAMEWORK: pytorch
MAX_JOBS: 1
run: |
docker exec -e NVTE_FRAMEWORK=pytorch -e NVTE_CUDA_ARCHS=${{ matrix.arch }} -e MAX_JOBS=2 builder \
bash -c 'pip install --no-build-isolation . -v --no-deps'
- name: 'Sanity check'
run: docker exec builder bash -c 'python3 tests/pytorch/test_sanity_import.py'
run: |
docker exec builder bash -c '\
ln -sf /usr/local/cuda/lib64/stubs/libcuda.so /usr/lib/x86_64-linux-gnu/libcuda.so.1 && \
python3 tests/pytorch/test_sanity_import.py \
'
jax:
name: 'JAX'
name: 'JAX (sm${{ matrix.arch }})'
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
arch: ["80", "90"]
container:
image: ghcr.io/nvidia/jax:jax-2026-07-21
options: --user root
Expand All @@ -107,17 +127,17 @@ jobs:
uses: actions/checkout@v3
with:
submodules: recursive
- name: ccache
uses: mozilla-actions/sccache-action@7d986dd989559c6ecdb630a3fd2557667be217ad
# sccache's nvcc wrapper breaks with the CUDA 13.x toolkit in this image
- name: 'Build'
run: |
NVTE_CCACHE_BIN=sccache NVTE_USE_CCACHE=1 pip install --no-build-isolation . -v
run: pip install --no-build-isolation . -v
env:
NVTE_FRAMEWORK: jax
MAX_JOBS: 1
SCCACHE_GHA_ENABLED: "true"
NVTE_CUDA_ARCHS: ${{ matrix.arch }}
MAX_JOBS: 2
- name: 'Sanity check'
run: python3 tests/jax/test_sanity_import.py
run: |
ln -sf /usr/local/cuda/lib64/stubs/libcuda.so /usr/lib/x86_64-linux-gnu/libcuda.so.1
python3 tests/jax/test_sanity_import.py
all:
name: 'All'
runs-on: ubuntu-latest
Expand Down Expand Up @@ -158,9 +178,13 @@ jobs:
pip install torch --no-cache-dir --index-url https://download.pytorch.org/whl/cu130
'
- name: 'Build'
run: docker exec builder bash -c 'pip install --no-cache-dir --no-build-isolation . -v --no-deps'
env:
NVTE_FRAMEWORK: all
MAX_JOBS: 1
run: |
docker exec -e NVTE_FRAMEWORK=all -e NVTE_CUDA_ARCHS=90 -e MAX_JOBS=2 builder \
bash -c 'pip install --no-cache-dir --no-build-isolation . -v --no-deps'
- name: 'Sanity check'
run: docker exec builder bash -c 'python3 tests/pytorch/test_sanity_import.py && python3 tests/jax/test_sanity_import.py'
run: |
docker exec builder bash -c '\
ln -sf /usr/local/cuda/lib64/stubs/libcuda.so /usr/lib/x86_64-linux-gnu/libcuda.so.1 && \
python3 tests/pytorch/test_sanity_import.py && \
python3 tests/jax/test_sanity_import.py \
'
Loading