minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[Test] Consolidate test cleanup and CI taxonomy (net -11.4K lines) (#37436)

Browse Source 
Co-authored-by: Mick Qian <mickqian@users.noreply.github.com>
This commit is contained in:
Xiaoyu Zhang
2026-09-07 15:13:59 +08:00
committed by GitHub
co-authored by Mick Qian
parent 6a1ff90f2d
commit 4d23a4fa6d
199 changed files with 1185 additions and 11812 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/pr-test-multimodal-gen.yml
+40 -40
View File
@@ -119,17 +119,16 @@ jobs:
timeout-minutes: 240
env:
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
PARTITION_PLAN_JSON: ${{ needs.compute-diffusion-partitions.outputs.plan-1gpu }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
DIFFUSION_PARTITION_ID: ${{ matrix.part }}
DIFFUSION_TOTAL_PARTITIONS: ${{ needs.compute-diffusion-partitions.outputs['partition-count-1gpu'] }}
DIFFUSION_PARTITION_PLAN_JSON: ${{ needs.compute-diffusion-partitions.outputs.plan-1gpu }}
SGLANG_DIFFUSION_ARTIFACT_DIR: ${{ github.workspace }}/diffusion-failures
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite 1-gpu \
--partition-id ${{ matrix.part }} \
--total-partitions ${{ needs.compute-diffusion-partitions.outputs['partition-count-1gpu'] }} \
--partition-plan-json "$PARTITION_PLAN_JSON" \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-1-gpu-h100 \
--timeout-per-file 14400
- name: Upload execution report
if: always()
@@ -192,13 +191,13 @@ jobs:
timeout-minutes: 120
env:
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
SGLANG_DIFFUSION_ARTIFACT_DIR: ${{ github.workspace }}/diffusion-failures
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite 1-gpu-5090 \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-1-gpu-5090 \
--timeout-per-file 7200
- name: Upload execution report
if: always()
@@ -261,13 +260,13 @@ jobs:
timeout-minutes: 60
env:
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
SGLANG_DIFFUSION_ARTIFACT_DIR: ${{ github.workspace }}/diffusion-bcg-artifacts
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite bcg-diffusion \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-bcg-1-gpu-h100 \
--timeout-per-file 3600
- name: Upload BCG diffusion artifacts
if: always()
@@ -329,17 +328,16 @@ jobs:
HF_TOKEN: ${{ secrets.SGLANG_DIFFUSION_CI_HF_TOKEN || secrets.HF_TOKEN }}
HUGGING_FACE_HUB_TOKEN: ${{ secrets.SGLANG_DIFFUSION_CI_HF_TOKEN || secrets.HF_TOKEN }}
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
PARTITION_PLAN_JSON: ${{ needs.compute-diffusion-partitions.outputs.plan-2gpu }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
DIFFUSION_PARTITION_ID: ${{ matrix.part }}
DIFFUSION_TOTAL_PARTITIONS: ${{ needs.compute-diffusion-partitions.outputs['partition-count-2gpu'] }}
DIFFUSION_PARTITION_PLAN_JSON: ${{ needs.compute-diffusion-partitions.outputs.plan-2gpu }}
SGLANG_DIFFUSION_ARTIFACT_DIR: ${{ github.workspace }}/diffusion-failures
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite 2-gpu \
--partition-id ${{ matrix.part }} \
--total-partitions ${{ needs.compute-diffusion-partitions.outputs['partition-count-2gpu'] }} \
--partition-plan-json "$PARTITION_PLAN_JSON" \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-2-gpu-h100 \
--timeout-per-file 14400
- name: Upload execution report
if: always()
@@ -402,12 +400,12 @@ jobs:
timeout-minutes: 240
env:
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite component-accuracy \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-component-2-gpu-h100 \
--timeout-per-file 14400
- uses: ./.github/actions/upload-cuda-coredumps
if: always()
@@ -453,13 +451,13 @@ jobs:
timeout-minutes: 240
env:
RUNAI_STREAMER_MEMORY_LIMIT: 0
CONTINUE_ON_ERROR_FLAG: ${{ inputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
DIFFUSION_CONTINUE_ON_ERROR: ${{ inputs.continue_on_error }}
SGLANG_DIFFUSION_ARTIFACT_DIR: ${{ github.workspace }}/diffusion-failures
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py \
--suite 1-gpu-b200 \
$CONTINUE_ON_ERROR_FLAG
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-1-gpu-b200 \
--timeout-per-file 14400
- name: Upload diffusion failure artifacts
if: always()
@@ -519,8 +517,10 @@ jobs:
- name: Run diffusion unit tests
timeout-minutes: 60
run: |
cd python
python3 sglang/multimodal_gen/test/run_suite.py --suite unit
python3 test/run_suite.py \
--hw cuda \
--suite base-b-test-diffusion-unit-1-gpu-h100 \
--timeout-per-file 3600
diffusion-coverage-check:
needs: [multimodal-gen-test-1-gpu, multimodal-gen-test-2-gpu]