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

[AMD][CI] Add GPT-OSS perf benchmarks to the ROCm 7.2 nightly (#34645)

Browse Source 
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Michael <michaelzhang-ai@users.noreply.github.com>
This commit is contained in:
Michael
2026-08-16 16:35:09 -07:00
committed by GitHub
co-authored by Cursor Agent Michael
parent f7cb328eb7
commit d91c3682b0
4 changed files with 283 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/nightly-test-amd-rocm720.yml
+30
View File
@@ -493,6 +493,7 @@ jobs:
- name: Accuracy Test ROCm 7.2 (8-GPU GPT-OSS)
timeout-minutes: 180
run: |
> github_summary.md # Clear summary file
bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
-e SGLANG_MOE_COPY_WEIGHT_VIEWS_BEFORE_H2D=1 \
-e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
@@ -500,6 +501,21 @@ jobs:
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
# Shares this job with the accuracy step above: it already serves the same
# two models, and an MI30x job spends ~49 min pulling the image and
# installing dependencies before it runs anything.
- name: Performance Test ROCm 7.2 (8-GPU GPT-OSS)
if: ${{ !cancelled() }}
timeout-minutes: 120
run: |
> github_summary.md # Clear summary file
bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
-e SGLANG_MOE_COPY_WEIGHT_VIEWS_BEFORE_H2D=1 \
-e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-gpt-oss --nightly --timeout-per-file 5400 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }} || TEST_EXIT_CODE=$?
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
nightly-accuracy-8-gpu-mi35x-rocm720:
if: (github.repository == 'sgl-project/sglang' || github.event_name == 'pull_request') && (!(inputs.job_filter || inputs.job_select) || (inputs.job_filter || inputs.job_select) == 'all' || contains(format(',{0},', inputs.job_filter || inputs.job_select), ',nightly-accuracy-8-gpu-mi35x-rocm720,'))
runs-on: linux-mi35x-gpu-8
@@ -528,12 +544,26 @@ jobs:
- name: Accuracy Test MI35x ROCm 7.2 (8-GPU GPT-OSS)
timeout-minutes: 180
run: |
> github_summary.md # Clear summary file
bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
-e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
python3 run_suite.py --hw amd --suite nightly-amd-8-gpu-mi35x --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}
# Shares this job with the accuracy step above rather than taking its own,
# so the container setup and the GPT-OSS weight cache are paid for once.
- name: Performance Test MI35x ROCm 7.2 (8-GPU GPT-OSS)
if: ${{ !cancelled() }}
timeout-minutes: 120
run: |
> github_summary.md # Clear summary file
bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
-e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-mi35x-gpt-oss --nightly --timeout-per-file 5400 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }} || TEST_EXIT_CODE=$?
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
exit ${TEST_EXIT_CODE:-0}
# ==============================================================================
# 8-GPU Grok1-INT4 (MI30x + MI35x)
# ==============================================================================