test(npu): remove obsolete npu pr nightly cases, move accuracy cases to full (#37990)

Co-authored-by: Sugar920 <Sugar920@users.noreply.github.com>
Co-authored-by: Claude Code <noreply@anthropic.com>
This commit is contained in:
Sugar920
2026-09-05 20:58:52 +08:00
committed by GitHub
co-authored by Sugar920 Claude Code
parent a18106bbc3
commit 4b802c052b
60 changed files with 283 additions and 3134 deletions
+3 -84
View File
@@ -122,22 +122,6 @@ jobs:
# Write to GITHUB_OUTPUT and print to the log in one command.
echo "run_start_metadata=${RUN_START_METADATA}" | tee -a $GITHUB_OUTPUT
nightly-1-npu-a2:
name: nightly-1-npu-a2
if: ${{ !cancelled() }}
needs: [set-image-config]
uses: ./.github/workflows/_npu-single-node-test-stage.yml
with:
runner: linux-aarch64-a2-4
test_type: 'perf'
test_suite: nightly-1-npu-a2
is_nightly_pipeline_job: true
skip_pr_test_health_check: 'true'
image: ${{ needs.set-image-config.outputs.image_a2 }}
install_sglang_deps: false
device_type_for_deps: '910b'
run_start_metadata: ${{ needs.set-image-config.outputs.run_start_metadata }}
nightly-1-npu-a3:
name: nightly-1-npu-a3
if: ${{ !cancelled() }}
@@ -247,22 +231,6 @@ jobs:
device_type_for_deps: 'a3'
run_start_metadata: ${{ needs.set-image-config.outputs.run_start_metadata }}
nightly-perf-8-npu-a3:
name: nightly-perf-8-npu-a3
if: ${{ !cancelled() }}
needs: [set-image-config]
uses: ./.github/workflows/_npu-single-node-test-stage.yml
with:
runner: linux-aarch64-a3-800t-8
test_type: 'perf'
test_suite: nightly-perf-8-npu-a3
is_nightly_pipeline_job: true
skip_pr_test_health_check: 'true'
image: ${{ needs.set-image-config.outputs.image_a3 }}
install_sglang_deps: false
device_type_for_deps: 'a3'
run_start_metadata: ${{ needs.set-image-config.outputs.run_start_metadata }}
nightly-perf-16-npu-a3:
name: nightly-perf-16-npu-a3
if: ${{ !cancelled() }}
@@ -298,22 +266,6 @@ jobs:
device_type_for_deps: 'a3'
run_start_metadata: ${{ needs.set-image-config.outputs.run_start_metadata }}
nightly-acc-4-npu-a3:
name: nightly-acc-4-npu-a3
if: ${{ !cancelled() }}
needs: [set-image-config]
uses: ./.github/workflows/_npu-single-node-test-stage.yml
with:
runner: linux-aarch64-a3-4-
test_type: 'accuracy'
test_suite: nightly-acc-4-npu-a3
is_nightly_pipeline_job: true
skip_pr_test_health_check: 'true'
image: ${{ needs.set-image-config.outputs.image_a3 }}
install_sglang_deps: false
device_type_for_deps: 'a3'
run_start_metadata: ${{ needs.set-image-config.outputs.run_start_metadata }}
nightly-acc-16-npu-a3:
name: nightly-acc-16-npu-a3
if: ${{ !cancelled() }}
@@ -333,7 +285,7 @@ jobs:
nightly-poc-multi-node-tests:
name: multi-node-poc
if: ${{ !cancelled() }}
needs: [set-image-config, nightly-perf-2-npu-a3, nightly-perf-4-npu-a3, nightly-perf-16-npu-a3, nightly-acc-2-npu-a3, nightly-acc-4-npu-a3, nightly-acc-16-npu-a3]
needs: [set-image-config, nightly-perf-2-npu-a3, nightly-perf-4-npu-a3, nightly-perf-16-npu-a3, nightly-acc-2-npu-a3, nightly-acc-16-npu-a3]
strategy:
fail-fast: false
max-parallel: 1
@@ -354,21 +306,6 @@ jobs:
test_case: test/registered/npu/accuracy/glm5_1/test_npu_glm5_1_w4a8_1p1d_32p_in64k_out1k_50ms_aime26.py
test_type: 'accuracy'
prefill_decode_deployment: 'separation'
# mimo_v2_flash performance tests
- name: test_npu_mimo_v2_flash_1p1d_12p_in16k_out1_ttft_5s
prefill_size: 1
decode_size: 1
router_size: 1
test_case: test/registered/npu/performance/mimo_v2_flash/test_npu_mimo_v2_flash_1p1d_12p_in16k_out1_ttft_5s.py
test_type: 'perf'
prefill_decode_deployment: 'separation'
- name: test_npu_mimo_v2_flash_1p1d_12p_in16k_out1k_tpot_20ms
prefill_size: 1
decode_size: 1
router_size: 1
test_case: test/registered/npu/performance/mimo_v2_flash/test_npu_mimo_v2_flash_1p1d_12p_in16k_out1k_tpot_20ms.py
test_type: 'perf'
prefill_decode_deployment: 'separation'
# deepseek_v4_flash performance tests
- name: deepseek_v4_flash_w8a8_1p1d_16p_in8k_out1k_50ms
prefill_size: 1
@@ -395,21 +332,12 @@ jobs:
nightly-poc-multi-node-mix-tests:
name: multi-node-mix-poc
if: ${{ !cancelled() }}
needs: [set-image-config, nightly-perf-2-npu-a3, nightly-perf-4-npu-a3, nightly-perf-16-npu-a3, nightly-acc-2-npu-a3, nightly-acc-4-npu-a3, nightly-acc-16-npu-a3, nightly-poc-multi-node-tests]
needs: [set-image-config, nightly-perf-2-npu-a3, nightly-perf-4-npu-a3, nightly-perf-16-npu-a3, nightly-acc-2-npu-a3, nightly-acc-16-npu-a3, nightly-poc-multi-node-tests]
strategy:
fail-fast: false
max-parallel: 1
matrix:
test_config:
# kimi_k2_6 performance tests
- name: kimi_k2_6_w4a8_16p_in64k_out1k_100ms
node_size: 2
test_case: test/registered/npu/performance/kimi_k2_6/test_npu_kimi_k2_6_w4a8_16p_in64k_out1k_100ms.py
test_type: 'perf'
- name: kimi_k2_6_w4a8_16p_in64k_out1k_100ms_aime25
node_size: 2
test_case: test/registered/npu/accuracy/kimi_k2_6/test_npu_kimi_k2_6_w4a8_16p_in64k_out1k_100ms_aime25.py
test_type: 'accuracy'
# glm_5_2 accuracy tests
- name: glm_5_2_w4a8_16p_gpqa
node_size: 2
@@ -436,7 +364,6 @@ jobs:
check-all-jobs:
if: ${{ !cancelled() }}
needs:
- nightly-1-npu-a2
- nightly-1-npu-a3
- nightly-2-npu-a3
- nightly-4-npu-a3
@@ -444,10 +371,8 @@ jobs:
- nightly-16-npu-a3
- nightly-perf-2-npu-a3
- nightly-perf-4-npu-a3
- nightly-perf-8-npu-a3
- nightly-perf-16-npu-a3
- nightly-acc-2-npu-a3
- nightly-acc-4-npu-a3
- nightly-acc-16-npu-a3
- nightly-poc-multi-node-tests
- nightly-poc-multi-node-mix-tests
@@ -455,7 +380,6 @@ jobs:
steps:
- name: Generate results table
run: |
single_result_a2="${{ needs.nightly-1-npu-a2.result }}"
multi_result="${{ needs.nightly-poc-multi-node-tests.result }}"
mix_result="${{ needs.nightly-poc-multi-node-mix-tests.result }}"
@@ -469,10 +393,8 @@ jobs:
"${{ needs.nightly-16-npu-a3.result }}" \
"${{ needs.nightly-perf-2-npu-a3.result }}" \
"${{ needs.nightly-perf-4-npu-a3.result }}" \
"${{ needs.nightly-perf-8-npu-a3.result }}" \
"${{ needs.nightly-perf-16-npu-a3.result }}" \
"${{ needs.nightly-acc-2-npu-a3.result }}" \
"${{ needs.nightly-acc-4-npu-a3.result }}" \
"${{ needs.nightly-acc-16-npu-a3.result }}"; do
if [ "${r}" != "success" ] && [ "${r}" != "skipped" ]; then
single_result="failure"
@@ -493,7 +415,6 @@ jobs:
echo "" >> $GITHUB_STEP_SUMMARY
echo "| Group | Status |" >> $GITHUB_STEP_SUMMARY
echo "|-------|--------|" >> $GITHUB_STEP_SUMMARY
echo "| nightly-1-npu-a2 | $(group_icon ${single_result_a2}) ${single_result_a2} |" >> $GITHUB_STEP_SUMMARY
for entry in \
"nightly-1-npu-a3:${{ needs.nightly-1-npu-a3.result }}" \
"nightly-2-npu-a3:${{ needs.nightly-2-npu-a3.result }}" \
@@ -502,10 +423,8 @@ jobs:
"nightly-16-npu-a3:${{ needs.nightly-16-npu-a3.result }}" \
"nightly-perf-2-npu-a3:${{ needs.nightly-perf-2-npu-a3.result }}" \
"nightly-perf-4-npu-a3:${{ needs.nightly-perf-4-npu-a3.result }}" \
"nightly-perf-8-npu-a3:${{ needs.nightly-perf-8-npu-a3.result }}" \
"nightly-perf-16-npu-a3:${{ needs.nightly-perf-16-npu-a3.result }}" \
"nightly-acc-2-npu-a3:${{ needs.nightly-acc-2-npu-a3.result }}" \
"nightly-acc-4-npu-a3:${{ needs.nightly-acc-4-npu-a3.result }}" \
"nightly-acc-16-npu-a3:${{ needs.nightly-acc-16-npu-a3.result }}"; do
suite="${entry%%:*}"
r="${entry##*:}"
@@ -516,7 +435,7 @@ jobs:
echo "" >> $GITHUB_STEP_SUMMARY
FAIL=0
for r in "${single_result_a2}" "${single_result}" "${multi_result}" "${mix_result}"; do
for r in "${single_result}" "${multi_result}" "${mix_result}"; do
if [ "${r}" != "success" ] && [ "${r}" != "skipped" ]; then FAIL=1; fi
done
exit $FAIL