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

xpu: record per-model metrics to jsonl for nightly dashboard (#36699)

Browse Source 
Co-authored-by: arathi-hlab <arathi-hlab@users.noreply.github.com>
This commit is contained in:
ashwini rathi
2026-09-01 10:20:56 +08:00
committed by GitHub
co-authored by arathi-hlab
parent f50b4ad7ae
commit 783af667fb
6 changed files with 481 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/nightly-test-intel.yml
+33 -3
View File
@@ -95,17 +95,27 @@ jobs:
- name: Nightly Test (1-GPU XPU)
timeout-minutes: 240
run: |
touch github_summary.md
touch github_summary.md nightly-xpu-1-gpu-metrics.jsonl
docker exec ci_sglang_xpu bash -c "
source /opt/venv/bin/activate &&
cd /sglang-checkout/test &&
OLMOCR_BENCH_DIR=/sglang-checkout/olmOCR-bench/bench_data \
GITHUB_STEP_SUMMARY=/sglang-checkout/github_summary.md \
SGLANG_TEST_METRICS_FILE=/sglang-checkout/nightly-xpu-1-gpu-metrics.jsonl \
python3 run_suite.py --hw xpu --suite nightly-xpu-1-gpu --nightly --timeout-per-file 7200 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }}
" || TEST_EXIT_CODE=$?
echo "$(<github_summary.md)" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
- name: Upload per-model metrics jsonl
if: always()
uses: actions/upload-artifact@v4
with:
name: nightly-xpu-1-gpu-metrics
path: nightly-xpu-1-gpu-metrics.jsonl
if-no-files-found: warn
retention-days: 30
- name: Cleanup container
if: always()
run: |
@@ -170,16 +180,26 @@ jobs:
- name: Nightly Test (2-GPU XPU)
timeout-minutes: 240
run: |
touch github_summary.md
touch github_summary.md nightly-xpu-2-gpu-metrics.jsonl
docker exec ci_sglang_xpu bash -c "
source /opt/venv/bin/activate &&
cd /sglang-checkout/test &&
GITHUB_STEP_SUMMARY=/sglang-checkout/github_summary.md \
SGLANG_TEST_METRICS_FILE=/sglang-checkout/nightly-xpu-2-gpu-metrics.jsonl \
python3 run_suite.py --hw xpu --suite nightly-xpu-2-gpu --nightly --timeout-per-file 7200 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }}
" || TEST_EXIT_CODE=$?
echo "$(<github_summary.md)" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
- name: Upload per-model metrics jsonl
if: always()
uses: actions/upload-artifact@v4
with:
name: nightly-xpu-2-gpu-metrics
path: nightly-xpu-2-gpu-metrics.jsonl
if-no-files-found: warn
retention-days: 30
- name: Cleanup container
if: always()
run: |
@@ -244,16 +264,26 @@ jobs:
- name: Nightly Test (4-GPU XPU)
timeout-minutes: 480
run: |
touch github_summary.md
touch github_summary.md nightly-xpu-4-gpu-metrics.jsonl
docker exec ci_sglang_xpu bash -c "
source /opt/venv/bin/activate &&
cd /sglang-checkout/test &&
GITHUB_STEP_SUMMARY=/sglang-checkout/github_summary.md \
SGLANG_TEST_METRICS_FILE=/sglang-checkout/nightly-xpu-4-gpu-metrics.jsonl \
python3 run_suite.py --hw xpu --suite nightly-xpu-4-gpu --nightly --timeout-per-file 7200 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }}
" || TEST_EXIT_CODE=$?
echo "$(<github_summary.md)" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
- name: Upload per-model metrics jsonl
if: always()
uses: actions/upload-artifact@v4
with:
name: nightly-xpu-4-gpu-metrics
path: nightly-xpu-4-gpu-metrics.jsonl
if-no-files-found: warn
retention-days: 30
- name: Cleanup container
if: always()
run: |