Skip to content
Run Float8 Tests

[float8nocompile] Add alternate Triton kernels for FP8 conversion which use atomic_max-based algo instead of reduction-based algo #3353

Sign in to view logs

[float8nocompile] Add alternate Triton kernels for FP8 conversion which use atomic_max-based algo instead of reduction-based algo

[float8nocompile] Add alternate Triton kernels for FP8 conversion which use atomic_max-based algo instead of reduction-based algo #3353

Workflow file for this run

.github/workflows/float8_test.yml at 8683228
name: Run Float8 Tests
on:
push:
branches:
- main
- 'gh/**'
pull_request:
branches:
- main
- 'gh/**'
concurrency:
group: float8_test-${{ github.workflow }}-${{ github.ref == 'refs/heads/main' && github.run_number || github.ref }}
cancel-in-progress: true
env:
HF_TOKEN: ${{ secrets.HF_TOKEN }}
jobs:
test:
strategy:
fail-fast: false
matrix:
include:
- name: SM-89
runs-on: linux.g6.4xlarge.experimental.nvidia.gpu
torch-spec: '--pre torch --index-url https://download.pytorch.org/whl/nightly/cu121'
gpu-arch-type: "cuda"
gpu-arch-version: "12.1"
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
with:
timeout: 60
runner: ${{ matrix.runs-on }}
gpu-arch-type: ${{ matrix.gpu-arch-type }}
gpu-arch-version: ${{ matrix.gpu-arch-version }}
script: |
conda create -n venv python=3.9 -y
conda activate venv
export PATH=/opt/rh/devtoolset-10/root/usr/bin/:$PATH
python -m pip install --upgrade pip
pip install ${{ matrix.torch-spec }}
pip install -r dev-requirements.txt
pip install .
pytest test/float8 --verbose -s