[AMD] Replace MI325 with MI300 CI Runners (#31409)
This commit is contained in:
Bingxu Chen
@@ -64,7 +64,7 @@ on:
|
||||
description: 'AMD runner pool to dispatch GPU jobs to'
|
||||
required: false
|
||||
type: choice
|
||||
default: mi325
|
||||
default: mi300
|
||||
options:
|
||||
- mi300
|
||||
- mi325
|
||||
@@ -80,6 +80,11 @@ on:
|
||||
required: false
|
||||
type: string
|
||||
default: ''
|
||||
runner_arch:
|
||||
description: 'AMD runner pool to dispatch GPU jobs to'
|
||||
required: false
|
||||
type: string
|
||||
default: mi300
|
||||
run_all_tests:
|
||||
description: "Run all tests (for releasing or testing purpose)"
|
||||
required: false
|
||||
@@ -217,13 +222,14 @@ jobs:
|
||||
uses: ./.github/workflows/pr-test-amd-extra.yml
|
||||
with:
|
||||
ref: ${{ inputs.pr_head_sha || inputs.ref || '' }}
|
||||
runner_arch: ${{ inputs.runner_arch || 'mi300' }}
|
||||
aiter_ref: ${{ inputs.aiter_ref }}
|
||||
continue_on_error: true
|
||||
secrets: inherit
|
||||
|
||||
# =============================================== sgl-kernel ====================================================
|
||||
sgl-kernel-unit-test-amd:
|
||||
name: ${{ format('sgl-kernel-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('sgl-kernel-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -235,7 +241,7 @@ jobs:
|
||||
needs.check-changes.outputs.sgl_kernel == 'true'
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -281,7 +287,7 @@ jobs:
|
||||
exit $failures
|
||||
|
||||
sgl-kernel-unit-test-2-gpu-amd:
|
||||
name: ${{ format('sgl-kernel-unit-test-2-gpu-amd (linux-{0}-2gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('sgl-kernel-unit-test-2-gpu-amd (linux-{0}-2gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -293,7 +299,7 @@ jobs:
|
||||
needs.check-changes.outputs.sgl_kernel == 'true'
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -322,7 +328,7 @@ jobs:
|
||||
# =============================================== primary ====================================================
|
||||
|
||||
stage-a-test-1-gpu-small-amd:
|
||||
name: ${{ format('stage-a-test-1-gpu-small-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('stage-a-test-1-gpu-small-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -334,7 +340,7 @@ jobs:
|
||||
((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -359,7 +365,7 @@ jobs:
|
||||
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-1-gpu-small-amd ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
jit-kernel-unit-test-amd:
|
||||
name: ${{ format('jit-kernel-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('jit-kernel-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -371,7 +377,7 @@ jobs:
|
||||
needs.check-changes.outputs.jit_kernel == 'true'
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -396,7 +402,7 @@ jobs:
|
||||
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite jit-kernel-unit-test-amd ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
jit-kernel-benchmark-test-amd:
|
||||
name: ${{ format('jit-kernel-benchmark-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('jit-kernel-benchmark-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -408,7 +414,7 @@ jobs:
|
||||
needs.check-changes.outputs.jit_kernel == 'true'
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -459,7 +465,7 @@ jobs:
|
||||
max-wait-minutes: '240'
|
||||
|
||||
stage-b-test-1-gpu-small-amd:
|
||||
name: ${{ format('stage-b-test-1-gpu-small-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
|
||||
name: ${{ format('stage-b-test-1-gpu-small-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, wait-for-stage-a-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -473,9 +479,10 @@ jobs:
|
||||
)
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 4
|
||||
matrix:
|
||||
part: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -494,12 +501,12 @@ jobs:
|
||||
run: bash scripts/ci/amd/amd_ci_install_dependency.sh
|
||||
|
||||
- name: Run test
|
||||
timeout-minutes: 60
|
||||
timeout-minutes: ${{ (inputs.runner_arch || 'mi300') == 'mi300' && 75 || 60 }}
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 14 --timeout-per-file 2400 ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
stage-b-test-1-gpu-small-amd-nondeterministic:
|
||||
name: ${{ format('stage-b-test-1-gpu-small-amd-nondeterministic (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('stage-b-test-1-gpu-small-amd-nondeterministic (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, wait-for-stage-a-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -511,7 +518,7 @@ jobs:
|
||||
((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -574,7 +581,7 @@ jobs:
|
||||
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-small-amd-mi35x ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
stage-b-test-1-gpu-large-amd:
|
||||
name: ${{ format('stage-b-test-1-gpu-large-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
|
||||
name: ${{ format('stage-b-test-1-gpu-large-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, wait-for-stage-a-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -588,9 +595,10 @@ jobs:
|
||||
)
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 1
|
||||
matrix:
|
||||
part: [0, 1, 2]
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -609,12 +617,12 @@ jobs:
|
||||
run: bash scripts/ci/amd/amd_ci_install_dependency.sh
|
||||
|
||||
- name: Run test
|
||||
timeout-minutes: 45
|
||||
timeout-minutes: ${{ (inputs.runner_arch || 'mi300') == 'mi300' && 60 || 45 }}
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-1-gpu-large-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 3 --timeout-per-file 2700 ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
stage-b-test-2-gpu-large-amd:
|
||||
name: ${{ format('stage-b-test-2-gpu-large-amd (linux-{0}-2gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
|
||||
name: ${{ format('stage-b-test-2-gpu-large-amd (linux-{0}-2gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, wait-for-stage-a-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -628,9 +636,10 @@ jobs:
|
||||
)
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 1
|
||||
matrix:
|
||||
part: [0, 1]
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -649,12 +658,12 @@ jobs:
|
||||
run: bash scripts/ci/amd/amd_ci_install_dependency.sh
|
||||
|
||||
- name: Run test
|
||||
timeout-minutes: 45
|
||||
timeout-minutes: ${{ (inputs.runner_arch || 'mi300') == 'mi300' && 120 || 45 }}
|
||||
run: |
|
||||
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' || '' }}
|
||||
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 ${{ (inputs.runner_arch || 'mi300') == 'mi300' && 5400 || 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 || 'mi325', matrix.part) }}
|
||||
name: ${{ format('multimodal-gen-test-1-gpu-amd (linux-{0}-1gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -668,9 +677,10 @@ jobs:
|
||||
)
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 1
|
||||
matrix:
|
||||
part: [0, 1, 2, 3]
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -699,16 +709,7 @@ jobs:
|
||||
|
||||
- name: Setup kernel caches
|
||||
run: |
|
||||
# Use the persistent /sgl-data directory (mounted from /home/runner/sgl-data)
|
||||
# This directory persists across container restarts on the self-hosted runner
|
||||
docker exec ci_sglang mkdir -p /sgl-data/aiter-kernels /sgl-data/miopen-cache /sgl-data/hf-cache/hub
|
||||
|
||||
# Clear pre-built AITER kernels from Docker image to avoid segfaults
|
||||
# The image may have stale/incompatible kernels at /sgl-workspace/aiter/aiter/jit/
|
||||
echo "Clearing pre-built AITER kernels from Docker image..."
|
||||
docker exec ci_sglang rm -rf /sgl-workspace/aiter/aiter/jit/*.so 2>/dev/null || true
|
||||
docker exec ci_sglang rm -rf /sgl-data/aiter-kernels/*.so 2>/dev/null || true
|
||||
echo "AITER kernels cleared - will be rebuilt on first use"
|
||||
docker exec ci_sglang mkdir -p /sgl-data/miopen-cache /sgl-data/hf-cache/hub
|
||||
|
||||
# Create persistent cache marker if /sgl-data is a real mount (not ephemeral)
|
||||
# This tells the test cleanup code to NOT delete downloaded models
|
||||
@@ -729,7 +730,7 @@ jobs:
|
||||
free -h
|
||||
echo ""
|
||||
echo "=== Disk Space ==="
|
||||
df -h /home/runner/sgl-data 2>/dev/null || df -h
|
||||
df -h /home/runner/sglang-data 2>/dev/null || df -h
|
||||
echo ""
|
||||
echo "=== HF Cache Directory Structure ==="
|
||||
docker exec ci_sglang ls -la /sgl-data/hf-cache/ 2>/dev/null || echo "HF cache dir not found"
|
||||
@@ -768,7 +769,6 @@ jobs:
|
||||
-e SGLANG_SKIP_CONSISTENCY=1 \
|
||||
-e SGLANG_TEST_NUM_INFERENCE_STEPS=5 \
|
||||
-e SGLANG_DIFFUSION_ARTIFACT_DIR=/sglang-checkout/diffusion-failures \
|
||||
-e AITER_JIT_DIR=/sgl-data/aiter-kernels \
|
||||
-e MIOPEN_USER_DB_PATH=/sgl-data/miopen-cache \
|
||||
-e HF_HUB_ENABLE_HF_TRANSFER=1 \
|
||||
-e HF_HUB_DISABLE_SYMLINKS_WARNING=1 \
|
||||
@@ -794,7 +794,7 @@ 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 || 'mi325', matrix.part) }}
|
||||
name: ${{ format('multimodal-gen-test-2-gpu-amd (linux-{0}-2gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -808,9 +808,10 @@ jobs:
|
||||
)
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 1
|
||||
matrix:
|
||||
part: [0, 1, 2] # 3 partitions: 2 parametrized + 1 standalone (single_test_file/test_disagg_server.py)
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-2gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -839,15 +840,7 @@ jobs:
|
||||
|
||||
- name: Setup kernel caches
|
||||
run: |
|
||||
# Use the persistent /sgl-data directory (mounted from /home/runner/sgl-data)
|
||||
docker exec ci_sglang mkdir -p /sgl-data/aiter-kernels /sgl-data/miopen-cache /sgl-data/hf-cache/hub
|
||||
|
||||
# Clear pre-built AITER kernels from Docker image to avoid segfaults
|
||||
# The image may have stale/incompatible kernels at /sgl-workspace/aiter/aiter/jit/
|
||||
echo "Clearing pre-built AITER kernels from Docker image..."
|
||||
docker exec ci_sglang rm -rf /sgl-workspace/aiter/aiter/jit/*.so 2>/dev/null || true
|
||||
docker exec ci_sglang rm -rf /sgl-data/aiter-kernels/*.so 2>/dev/null || true
|
||||
echo "AITER kernels cleared - will be rebuilt on first use"
|
||||
docker exec ci_sglang mkdir -p /sgl-data/miopen-cache /sgl-data/hf-cache/hub
|
||||
|
||||
# Create persistent cache marker if /sgl-data is a real mount (not ephemeral)
|
||||
# This tells the test cleanup code to NOT delete downloaded models
|
||||
@@ -868,7 +861,7 @@ jobs:
|
||||
free -h
|
||||
echo ""
|
||||
echo "=== Disk Space ==="
|
||||
df -h /home/runner/sgl-data 2>/dev/null || df -h
|
||||
df -h /home/runner/sglang-data 2>/dev/null || df -h
|
||||
echo ""
|
||||
echo "=== HF Cache Directory Structure ==="
|
||||
docker exec ci_sglang ls -la /sgl-data/hf-cache/ 2>/dev/null || echo "HF cache dir not found"
|
||||
@@ -890,7 +883,7 @@ jobs:
|
||||
docker exec ci_sglang rocm-smi --showmeminfo vram 2>/dev/null || echo "rocm-smi not available"
|
||||
|
||||
- name: Run diffusion server tests (2-GPU)
|
||||
timeout-minutes: 150
|
||||
timeout-minutes: ${{ (inputs.runner_arch || 'mi300') == 'mi300' && 180 || 150 }}
|
||||
run: |
|
||||
# AMD CI: All 2-GPU tests including LoRA
|
||||
# Tests: T2V, T2I, I2V, LoRA
|
||||
@@ -907,7 +900,6 @@ jobs:
|
||||
-e SGLANG_SKIP_CONSISTENCY=1 \
|
||||
-e SGLANG_TEST_NUM_INFERENCE_STEPS=5 \
|
||||
-e SGLANG_DIFFUSION_ARTIFACT_DIR=/sglang-checkout/diffusion-failures \
|
||||
-e AITER_JIT_DIR=/sgl-data/aiter-kernels \
|
||||
-e MIOPEN_USER_DB_PATH=/sgl-data/miopen-cache \
|
||||
-e HF_HUB_ENABLE_HF_TRANSFER=1 \
|
||||
-e HF_HUB_DISABLE_SYMLINKS_WARNING=1 \
|
||||
@@ -937,7 +929,7 @@ jobs:
|
||||
# that don't require NVIDIA hardware, so they should run on AMD too. This
|
||||
# closes the AMD coverage gap the dashboard surfaces for these tests.
|
||||
multimodal-gen-unit-test-amd:
|
||||
name: ${{ format('multimodal-gen-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi325') }}
|
||||
name: ${{ format('multimodal-gen-unit-test-amd (linux-{0}-1gpu-sglang)', inputs.runner_arch || 'mi300') }}
|
||||
needs: [check-changes, call-gate]
|
||||
if: |
|
||||
always() && !cancelled() &&
|
||||
@@ -949,7 +941,7 @@ jobs:
|
||||
needs.check-changes.outputs.multimodal_gen == 'true'
|
||||
)
|
||||
)
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-1gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -983,7 +975,6 @@ jobs:
|
||||
# memory-format behavior that the ROCm conv path doesn't reproduce
|
||||
# (CUDA-specific). The rest of the unit suite is portable.
|
||||
docker exec \
|
||||
-e AITER_JIT_DIR=/sgl-data/aiter-kernels \
|
||||
-e MIOPEN_USER_DB_PATH=/sgl-data/miopen-cache \
|
||||
-w /sglang-checkout/python \
|
||||
ci_sglang python3 sglang/multimodal_gen/test/run_suite.py \
|
||||
@@ -1019,7 +1010,7 @@ jobs:
|
||||
max-wait-minutes: '480'
|
||||
|
||||
stage-c-test-4-gpu-amd:
|
||||
name: ${{ format('stage-c-test-4-gpu-amd (linux-{0}-4gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
|
||||
name: ${{ format('stage-c-test-4-gpu-amd (linux-{0}-4gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, call-gate, wait-for-stage-b-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -1035,7 +1026,7 @@ jobs:
|
||||
fail-fast: false
|
||||
matrix:
|
||||
part: [0]
|
||||
runs-on: ${{ format('linux-{0}-4gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-4gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
@@ -1075,7 +1066,7 @@ jobs:
|
||||
${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
|
||||
|
||||
stage-c-test-large-8-gpu-amd:
|
||||
name: ${{ format('stage-c-test-large-8-gpu-amd (linux-{0}-8gpu-sglang, {1})', inputs.runner_arch || 'mi325', matrix.part) }}
|
||||
name: ${{ format('stage-c-test-large-8-gpu-amd (linux-{0}-8gpu-sglang, {1})', inputs.runner_arch || 'mi300', matrix.part) }}
|
||||
needs: [check-changes, call-gate, wait-for-stage-b-amd]
|
||||
if: |
|
||||
always() &&
|
||||
@@ -1088,12 +1079,13 @@ jobs:
|
||||
)
|
||||
)
|
||||
env:
|
||||
RUNNER_LABELS: ${{ format('linux-{0}-8gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
RUNNER_LABELS: ${{ format('linux-{0}-8gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
strategy:
|
||||
fail-fast: false
|
||||
max-parallel: 1
|
||||
matrix:
|
||||
part: [0, 1, 2, 3]
|
||||
runs-on: ${{ format('linux-{0}-8gpu-sglang', inputs.runner_arch || 'mi325') }}
|
||||
runs-on: ${{ format('linux-{0}-8gpu-sglang', inputs.runner_arch || 'mi300') }}
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
|
||||
Reference in New Issue
Block a user