76 lines
3.0 KiB
YAML
76 lines
3.0 KiB
YAML
# Weekly CUDA tests. Same shape as nightly-test-nvidia.yml (see its header); a
|
|
# test joins by declaring `stage="weekly"` plus a runner_config.
|
|
name: Weekly Test (Nvidia)
|
|
|
|
on:
|
|
schedule:
|
|
- cron: '0 0 * * 0' # Run every Sunday at midnight UTC
|
|
workflow_dispatch:
|
|
inputs:
|
|
runner_filter:
|
|
# A free-text field rather than a choice: `on:` blocks take no
|
|
# expressions, so an options list would be a second copy of the matrix
|
|
# below to keep in sync. Matched in _pr-test-stage.yml.
|
|
description: 'runner_config to run alone (empty or "all" runs every one)'
|
|
required: false
|
|
type: string
|
|
default: 'all'
|
|
full_parallel:
|
|
description: 'Run all shards of a job at once (faster, but competes with per-commit CI for machines). Off by default: one shard at a time.'
|
|
required: false
|
|
type: boolean
|
|
default: false
|
|
|
|
concurrency:
|
|
group: weekly-test-nvidia-${{ github.ref }}
|
|
cancel-in-progress: true
|
|
|
|
permissions:
|
|
actions: write
|
|
contents: read
|
|
issues: read
|
|
pull-requests: read
|
|
|
|
jobs:
|
|
# run_all_tests skips the paths-filter, so main_package is 'true' and
|
|
# sgl_kernel stays empty: every test runs, and no job waits on a wheel this
|
|
# workflow never builds.
|
|
check-changes:
|
|
uses: ./.github/workflows/_pr-test-check-changes.yml
|
|
with:
|
|
pr_test_yml: '.github/workflows/weekly-test-nvidia.yml'
|
|
run_all_tests: true
|
|
force_continue_on_error: true
|
|
secrets: inherit
|
|
|
|
# One row per machine type; timeouts mirror the same runner's nightly job.
|
|
# Adding a machine is one row here, and adding a test to an existing machine
|
|
# is none -- `stage="weekly"` + `runner_config=` on the test is enough.
|
|
weekly-test:
|
|
# Without this, the matrix default names the job after every value in the
|
|
# row -- "weekly-test (1-gpu-large, 120, 180)".
|
|
name: weekly-test-${{ matrix.runner_config }}
|
|
needs: check-changes
|
|
if: github.repository == 'sgl-project/sglang'
|
|
strategy:
|
|
fail-fast: false
|
|
matrix:
|
|
include:
|
|
- { runner_config: 1-gpu-large, run_timeout: '120', job_timeout: '180' }
|
|
- { runner_config: 2-gpu-large, run_timeout: '240', job_timeout: '300' }
|
|
- { runner_config: 4-gpu-h100, run_timeout: '120', job_timeout: '180' }
|
|
- { runner_config: 4-gpu-b200, run_timeout: '150', job_timeout: '210' }
|
|
- { runner_config: 8-gpu-h200, run_timeout: '240', job_timeout: '300' }
|
|
- { runner_config: 8-gpu-b200, run_timeout: '360', job_timeout: '420' }
|
|
uses: ./.github/workflows/_pr-test-stage.yml
|
|
with:
|
|
self_name: weekly-test-${{ matrix.runner_config }}
|
|
runner_config: ${{ matrix.runner_config }}
|
|
check_changes: ${{ toJson(needs.check-changes.outputs) }}
|
|
caller_inputs: ${{ toJson(inputs) }}
|
|
partitions: ${{ needs.check-changes.outputs.partitions }}
|
|
run_timeout_minutes: ${{ matrix.run_timeout }}
|
|
job_timeout_minutes: ${{ matrix.job_timeout }}
|
|
scheduled: true
|
|
secrets: inherit
|