minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 10 Packages Projects Releases Wiki Activity

[AMD] Route PR multimodal tests to MI325 (#24614)

Browse Source 
Co-authored-by: bingxche <bingxche@amd.com>
This commit is contained in:
YC Yen-Ching Tseng
2026-05-07 23:26:05 +08:00
committed by GitHub
co-authored by bingxche
parent 9c41b1058f
commit c18879054f
1 changed files with 4 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/pr-test-amd.yml
+4 -6
View File
@@ -579,7 +579,7 @@ jobs:
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-2-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 2700 ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
multimodal-gen-test-1-gpu-amd:
name: ${{ format('multimodal-gen-test-1-gpu-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || (github.event_name == 'pull_request' && 'mi300' || 'mi325'), matrix.part) }}
name: ${{ format('multimodal-gen-test-1-gpu-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
needs: [check-changes, call-gate]
if: |
always() && !cancelled() &&
@@ -593,10 +593,9 @@ jobs:
)
strategy:
fail-fast: false
max-parallel: 1 # Run one at a time to avoid eviction from resource exhaustion during AITER kernel JIT
matrix:
part: [0, 1, 2, 3]
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || (github.event_name == 'pull_request' && 'mi300' || 'mi325')) }}
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
steps:
- name: Checkout code
uses: actions/checkout@v4
@@ -720,7 +719,7 @@ jobs:
retention-days: 7
multimodal-gen-test-2-gpu-amd:
name: ${{ format('multimodal-gen-test-2-gpu-amd (linux-{0}-2gpu-sglang, {1})', inputs.runner_arch || (github.event_name == 'pull_request' && 'mi300' || 'mi325'), matrix.part) }}
name: ${{ format('multimodal-gen-test-2-gpu-amd (linux-{0}-2gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
needs: [check-changes, call-gate]
if: |
always() && !cancelled() &&
@@ -734,10 +733,9 @@ jobs:
)
strategy:
fail-fast: false
max-parallel: 1 # Run one at a time to avoid eviction from resource exhaustion during AITER kernel JIT
matrix:
part: [0, 1, 2] # 3 partitions: 2 parametrized + 1 standalone (test_disagg_server.py)
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || (github.event_name == 'pull_request' && 'mi300' || 'mi325')) }}
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi325') }}
steps:
- name: Checkout code
uses: actions/checkout@v4