[AMD] [CI] Register MI35x GSM8K nightly tests (#26478)
Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Bingxu Chen
@@ -30,12 +30,14 @@ on:
|
||||
# 1-GPU Unit Tests (MI30x + MI35x)
|
||||
- nightly-test-1-gpu-unit-rocm720
|
||||
- nightly-test-1-gpu-mi35x-rocm720
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
- nightly-accuracy-2-gpu-rocm720
|
||||
- nightly-accuracy-2-gpu-vlm-rocm720
|
||||
- nightly-perf-2-gpu-text-rocm720
|
||||
- nightly-perf-2-gpu-vlm-rocm720
|
||||
- nightly-4-gpu-rocm720
|
||||
# 2-GPU GLM-5.1-MXFP4 (MI35x only)
|
||||
- nightly-2-gpu-mi35x-glm51-mxfp4-rocm720
|
||||
# 8-GPU GPT-OSS (MI30x mixes Grok1-FP8; MI35x mixes Qwen3-Coder-Next)
|
||||
- nightly-accuracy-8-gpu-rocm720
|
||||
- nightly-accuracy-8-gpu-mi35x-rocm720
|
||||
@@ -55,10 +57,11 @@ on:
|
||||
- nightly-accuracy-8-gpu-mi35x-deepseek-v32-mtp-rocm720
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-basic-rocm720
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-mtp-rocm720
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-kv-fp8-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-ar-fusion-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-hicache-rocm720
|
||||
# 8-GPU DeepSeek-V4 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-v4-flash-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-v4-pro-rocm720
|
||||
@@ -191,7 +194,7 @@ jobs:
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
# ==============================================================================
|
||||
|
||||
nightly-accuracy-2-gpu-rocm720:
|
||||
@@ -224,6 +227,41 @@ jobs:
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
nightly-2-gpu-mi35x-glm51-mxfp4-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-2-gpu-mi35x-glm51-mxfp4-rocm720,'))
|
||||
runs-on: linux-mi35x-gpu-2
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
ref: ${{ inputs.ref || github.sha }}
|
||||
|
||||
- name: Ensure VRAM is clear
|
||||
run: bash scripts/ci/amd/ensure_vram_clear.sh rocm
|
||||
|
||||
- name: Setup docker (ROCm 7.2)
|
||||
run: |
|
||||
touch github_summary.md
|
||||
bash scripts/ci/amd/amd_ci_start_container.sh --rocm-version rocm720
|
||||
env:
|
||||
GITHUB_WORKSPACE: ${{ github.workspace }}
|
||||
|
||||
- name: Install dependencies
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_install_dependency.sh --skip-test-time-deps
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install git+https://github.com/huggingface/transformers.git@96f807a33b75
|
||||
|
||||
- name: Accuracy Test MI35x ROCm 7.2 (2-GPU GLM-5.1-MXFP4 GSM8K)
|
||||
timeout-minutes: 120
|
||||
run: |
|
||||
> github_summary.md
|
||||
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-2-gpu-mi35x-glm51-mxfp4 --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}
|
||||
|
||||
nightly-accuracy-2-gpu-vlm-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-2-gpu-vlm-rocm720,'))
|
||||
runs-on: linux-mi325-2gpu-sglang
|
||||
@@ -936,7 +974,7 @@ jobs:
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
# ==============================================================================
|
||||
|
||||
nightly-8-gpu-mi35x-deepseek-r1-mxfp4-rocm720:
|
||||
@@ -1077,6 +1115,40 @@ jobs:
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
nightly-8-gpu-mi35x-deepseek-r1-hicache-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-8-gpu-mi35x-deepseek-r1-hicache-rocm720,'))
|
||||
runs-on: linux-mi35x-gpu-8
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
ref: ${{ inputs.ref || github.sha }}
|
||||
|
||||
- name: Ensure VRAM is clear
|
||||
run: bash scripts/ci/amd/ensure_vram_clear.sh rocm
|
||||
|
||||
- name: Setup docker (ROCm 7.2)
|
||||
run: |
|
||||
touch github_summary.md
|
||||
bash scripts/ci/amd/amd_ci_start_container.sh --rocm-version rocm720
|
||||
env:
|
||||
GITHUB_WORKSPACE: ${{ github.workspace }}
|
||||
|
||||
- name: Install dependencies
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_install_dependency.sh --skip-test-time-deps
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
|
||||
|
||||
- name: Accuracy Test MI35x ROCm 7.2 (8-GPU DeepSeek-R1 HiCache)
|
||||
timeout-minutes: 180
|
||||
run: |
|
||||
> github_summary.md
|
||||
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-deepseek-r1-hicache --nightly --timeout-per-file 3600 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }} || TEST_EXIT_CODE=$?
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 8-GPU DeepSeek-V4 (MI35x only)
|
||||
# ==============================================================================
|
||||
@@ -1693,12 +1765,13 @@ jobs:
|
||||
# 1-GPU Unit Tests (MI30x + MI35x)
|
||||
- nightly-test-1-gpu-unit-rocm720
|
||||
- nightly-test-1-gpu-mi35x-rocm720
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
- nightly-accuracy-2-gpu-rocm720
|
||||
- nightly-accuracy-2-gpu-vlm-rocm720
|
||||
- nightly-perf-2-gpu-text-rocm720
|
||||
- nightly-perf-2-gpu-vlm-rocm720
|
||||
- nightly-4-gpu-rocm720
|
||||
- nightly-2-gpu-mi35x-glm51-mxfp4-rocm720
|
||||
# 8-GPU GPT-OSS (MI30x mixes Grok1-FP8; MI35x mixes Qwen3-Coder-Next)
|
||||
- nightly-accuracy-8-gpu-rocm720
|
||||
- nightly-accuracy-8-gpu-mi35x-rocm720
|
||||
@@ -1718,10 +1791,11 @@ jobs:
|
||||
- nightly-accuracy-8-gpu-mi35x-deepseek-v32-mtp-rocm720
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-basic-rocm720
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-mtp-rocm720
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-kv-fp8-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-ar-fusion-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-hicache-rocm720
|
||||
# 8-GPU DeepSeek-V4 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-v4-flash-rocm720
|
||||
- nightly-8-gpu-mi35x-deepseek-v4-pro-rocm720
|
||||
|
||||
@@ -30,12 +30,14 @@ on:
|
||||
# 1-GPU Unit Tests (MI30x + MI35x)
|
||||
- nightly-test-1-gpu-unit
|
||||
- nightly-test-1-gpu-mi35x
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
- nightly-accuracy-2-gpu
|
||||
- nightly-accuracy-2-gpu-vlm
|
||||
- nightly-perf-2-gpu-text
|
||||
- nightly-perf-2-gpu-vlm
|
||||
- nightly-4-gpu
|
||||
# 2-GPU GLM-5.1-MXFP4 (MI35x only)
|
||||
- nightly-2-gpu-mi35x-glm51-mxfp4
|
||||
# 8-GPU GPT-OSS (MI30x mixes Grok1-FP8; MI35x mixes Qwen3-Coder-Next)
|
||||
- nightly-accuracy-8-gpu
|
||||
- nightly-accuracy-8-gpu-mi35x
|
||||
@@ -55,10 +57,11 @@ on:
|
||||
- nightly-accuracy-8-gpu-mi35x-deepseek-v32-mtp
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-basic
|
||||
- nightly-perf-8-gpu-mi35x-deepseek-v32-mtp
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-kv-fp8
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-ar-fusion
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-hicache
|
||||
# 8-GPU Kimi-K2.6 (MI30x + MI35x)
|
||||
- nightly-8-gpu-kimi-k26
|
||||
- nightly-8-gpu-mi35x-kimi-k26
|
||||
@@ -192,7 +195,7 @@ jobs:
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
# ==============================================================================
|
||||
|
||||
nightly-accuracy-2-gpu:
|
||||
@@ -226,6 +229,41 @@ jobs:
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
nightly-2-gpu-mi35x-glm51-mxfp4:
|
||||
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-2-gpu-mi35x-glm51-mxfp4,'))
|
||||
runs-on: linux-mi35x-gpu-2
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
ref: ${{ inputs.ref || github.sha }}
|
||||
|
||||
- name: Ensure VRAM is clear
|
||||
run: bash scripts/ci/amd/ensure_vram_clear.sh rocm
|
||||
|
||||
- name: Setup docker
|
||||
run: |
|
||||
touch github_summary.md
|
||||
bash scripts/ci/amd/amd_ci_start_container.sh
|
||||
env:
|
||||
GITHUB_WORKSPACE: ${{ github.workspace }}
|
||||
|
||||
- name: Install dependencies
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_install_dependency.sh
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install git+https://github.com/huggingface/transformers.git@96f807a33b75
|
||||
|
||||
- name: Accuracy Test MI35x (2-GPU GLM-5.1-MXFP4 GSM8K)
|
||||
timeout-minutes: 120
|
||||
run: |
|
||||
> github_summary.md
|
||||
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-2-gpu-mi35x-glm51-mxfp4 --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}
|
||||
|
||||
nightly-accuracy-2-gpu-vlm:
|
||||
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-2-gpu-vlm,'))
|
||||
runs-on: linux-mi325-2gpu-sglang
|
||||
@@ -941,7 +979,7 @@ jobs:
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
# ==============================================================================
|
||||
|
||||
nightly-8-gpu-mi35x-deepseek-r1-mxfp4:
|
||||
@@ -1082,6 +1120,40 @@ jobs:
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
nightly-8-gpu-mi35x-deepseek-r1-hicache:
|
||||
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-8-gpu-mi35x-deepseek-r1-hicache,'))
|
||||
runs-on: linux-mi35x-gpu-8
|
||||
steps:
|
||||
- name: Checkout code
|
||||
uses: actions/checkout@v4
|
||||
with:
|
||||
ref: ${{ inputs.ref || github.sha }}
|
||||
|
||||
- name: Ensure VRAM is clear
|
||||
run: bash scripts/ci/amd/ensure_vram_clear.sh rocm
|
||||
|
||||
- name: Setup docker
|
||||
run: |
|
||||
touch github_summary.md
|
||||
bash scripts/ci/amd/amd_ci_start_container.sh
|
||||
env:
|
||||
GITHUB_WORKSPACE: ${{ github.workspace }}
|
||||
|
||||
- name: Install dependencies
|
||||
run: |
|
||||
bash scripts/ci/amd/amd_ci_install_dependency.sh
|
||||
bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
|
||||
|
||||
- name: Accuracy Test MI35x (8-GPU DeepSeek-R1 HiCache)
|
||||
timeout-minutes: 180
|
||||
run: |
|
||||
> github_summary.md
|
||||
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-deepseek-r1-hicache --nightly --timeout-per-file 3600 ${{ (github.event_name == 'schedule' || inputs.continue_on_error) && '--continue-on-error' || '' }} || TEST_EXIT_CODE=$?
|
||||
echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
|
||||
exit ${TEST_EXIT_CODE:-0}
|
||||
|
||||
# ==============================================================================
|
||||
# 8-GPU Kimi-K2.6 (MI30x + MI35x)
|
||||
# ==============================================================================
|
||||
@@ -1569,12 +1641,13 @@ jobs:
|
||||
# 1-GPU Unit Tests (MI30x + MI35x)
|
||||
- nightly-test-1-gpu-unit
|
||||
- nightly-test-1-gpu-mi35x
|
||||
# 2-GPU and 4-GPU Tests (MI30x)
|
||||
# 2-GPU and 4-GPU Tests (MI30x + MI35x)
|
||||
- nightly-accuracy-2-gpu
|
||||
- nightly-accuracy-2-gpu-vlm
|
||||
# - nightly-perf-2-gpu-text # excluded: perf failures don't block CI
|
||||
# - nightly-perf-2-gpu-vlm # excluded: perf failures don't block CI
|
||||
- nightly-4-gpu
|
||||
- nightly-2-gpu-mi35x-glm51-mxfp4
|
||||
# 8-GPU GPT-OSS (MI30x mixes Grok1-FP8; MI35x mixes Qwen3-Coder-Next)
|
||||
- nightly-accuracy-8-gpu
|
||||
- nightly-accuracy-8-gpu-mi35x
|
||||
@@ -1594,10 +1667,11 @@ jobs:
|
||||
- nightly-accuracy-8-gpu-mi35x-deepseek-v32-mtp
|
||||
# - nightly-perf-8-gpu-mi35x-deepseek-v32-basic # excluded: perf failures don't block CI
|
||||
# - nightly-perf-8-gpu-mi35x-deepseek-v32-mtp # excluded: perf failures don't block CI
|
||||
# 8-GPU DeepSeek-R1-MXFP4 (MI35x only)
|
||||
# 8-GPU DeepSeek-R1 (MI35x only)
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-kv-fp8
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-mxfp4-ar-fusion
|
||||
- nightly-8-gpu-mi35x-deepseek-r1-hicache
|
||||
# 8-GPU Kimi-K2.6 (MI30x + MI35x)
|
||||
- nightly-8-gpu-kimi-k26
|
||||
- nightly-8-gpu-mi35x-kimi-k26
|
||||
|
||||
Reference in New Issue
Block a user