ci: remove IS_BLACKWELL env var; auto-detect Blackwell (#21118)

This commit is contained in:
Lianmin Zheng
2026-03-22 23:44:48 -07:00
committed by GitHub
parent 54fdb357d3
commit 7757a9ddd0
6 changed files with 14 additions and 17 deletions

View File

@@ -242,7 +242,7 @@ jobs:
- name: Install dependencies
run: |
IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
bash scripts/ci/cuda/ci_install_dependency.sh
- name: Run common 8-GPU model tests
if: always()
@@ -253,7 +253,7 @@ jobs:
GPU_CONFIG: "8-gpu-b200"
run: |
cd test
IS_BLACKWELL=1 python3 run_suite.py --hw cuda --suite nightly-8-gpu-common --nightly --timeout-per-file=12000 --continue-on-error --auto-partition-id=${{ matrix.partition }} --auto-partition-size=4
python3 run_suite.py --hw cuda --suite nightly-8-gpu-common --nightly --timeout-per-file=12000 --continue-on-error --auto-partition-id=${{ matrix.partition }} --auto-partition-size=4
- name: Publish traces to storage repo
if: always()
@@ -570,7 +570,7 @@ jobs:
- name: Install dependencies
run: |
IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
bash scripts/ci/cuda/ci_install_dependency.sh
- name: Run test
timeout-minutes: 300
@@ -599,7 +599,7 @@ jobs:
- name: Install dependencies
run: |
IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
bash scripts/ci/cuda/ci_install_dependency.sh
- name: Run test
timeout-minutes: 120

View File

@@ -949,8 +949,6 @@ jobs:
)
runs-on: 1-gpu-5090
timeout-minutes: 240
env:
IS_BLACKWELL: "1"
strategy:
fail-fast: false
max-parallel: 8
@@ -1148,7 +1146,7 @@ jobs:
- name: Install dependencies
timeout-minutes: 20
run: |
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
- name: Run test
timeout-minutes: 30
@@ -1156,12 +1154,12 @@ jobs:
CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
run: |
cd test
IS_BLACKWELL=1 python3 run_suite.py --hw cuda --suite stage-b-test-4-gpu-b200 $CONTINUE_ON_ERROR_FLAG
python3 run_suite.py --hw cuda --suite stage-b-test-4-gpu-b200 $CONTINUE_ON_ERROR_FLAG
- name: Run FA4 jit_kernel tests (SM100+)
timeout-minutes: 10
run: |
IS_BLACKWELL=1 python3 -m pytest -q python/sglang/jit_kernel/tests/test_flash_attention_4.py
python3 -m pytest -q python/sglang/jit_kernel/tests/test_flash_attention_4.py
- uses: ./.github/actions/upload-cuda-coredumps
if: always()
@@ -1665,7 +1663,7 @@ jobs:
- name: Install dependencies
timeout-minutes: 20
run: |
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
- name: Run test
timeout-minutes: 30
@@ -1673,7 +1671,7 @@ jobs:
CONTINUE_ON_ERROR_FLAG: ${{ needs.check-changes.outputs.continue_on_error == 'true' && '--continue-on-error' || '' }}
run: |
cd test
IS_BLACKWELL=1 python3 run_suite.py --hw cuda --suite stage-c-test-4-gpu-b200 --auto-partition-id ${{ matrix.part }} --auto-partition-size 4 --timeout-per-file 1800 $CONTINUE_ON_ERROR_FLAG
python3 run_suite.py --hw cuda --suite stage-c-test-4-gpu-b200 --auto-partition-id ${{ matrix.part }} --auto-partition-size 4 --timeout-per-file 1800 $CONTINUE_ON_ERROR_FLAG
- uses: ./.github/actions/upload-cuda-coredumps
if: always()
@@ -1719,7 +1717,7 @@ jobs:
# - name: Install dependencies
# timeout-minutes: 20
# run: |
# CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 GRACE_BLACKWELL=1 bash scripts/ci/cuda/ci_install_deepep.sh
# CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} GRACE_BLACKWELL=1 bash scripts/ci/cuda/ci_install_deepep.sh
#
# - name: Run test
# timeout-minutes: 45

View File

@@ -48,7 +48,6 @@ jobs:
timeout-minutes: 120
env:
RUNNER_LABELS: ${{ inputs.runner_label }}
IS_BLACKWELL: ${{ (inputs.runner_label == '1-gpu-5090' || contains(inputs.runner_label, 'b200')) && '1' || '0' }}
SGLANG_CI_RDMA_ALL_DEVICES: ${{ inputs.runner_label == '8-gpu-h20' && 'mlx5_1,mlx5_2,mlx5_3,mlx5_4' || '' }}
steps:
- name: Checkout code

View File

@@ -194,7 +194,6 @@ class Envs:
SGLANG_TEST_STUCK_SCHEDULER_INIT = EnvFloat(0)
SGLANG_TEST_STUCK_TOKENIZER = EnvFloat(0)
SGLANG_TEST_CRASH_AFTER_STREAM_OUTPUTS = EnvInt(0)
IS_BLACKWELL = EnvBool(False)
IS_H200 = EnvBool(False)
SGLANG_SET_CPU_AFFINITY = EnvBool(False)
SGLANG_PROFILE_WITH_STACK = EnvBool(True)

View File

@@ -37,6 +37,7 @@ from sglang.srt.environ import envs
from sglang.srt.utils import (
get_bool_env_var,
get_device,
is_blackwell,
is_cuda,
is_xpu,
kill_process_tree,
@@ -197,8 +198,8 @@ def is_in_amd_ci():
def is_blackwell_system():
"""Return whether it is running on a Blackwell (B200) system."""
return envs.IS_BLACKWELL.get()
"""Same CUDA capability + toolkit semantics as ``sglang.srt.utils.is_blackwell``."""
return is_blackwell()
def is_h200_system():

View File

@@ -36,7 +36,7 @@ HW_MAPPING = {
# Regex to detect run_suite.py invocations from pr-test.yml.
# Uses lookaheads so --hw and --suite can appear in any order.
_RUN_SUITE_RE = re.compile(
r"^\s*(?:\w+=\S+\s+)?" # optional env prefix (e.g. IS_BLACKWELL=1)
r"^\s*(?:\w+=\S+\s+)?" # optional env prefix (e.g. VAR=value)
r"python3\s+run_suite\.py\b"
r"(?=.*\s--hw\s+(?P<hw>\S+))"
r"(?=.*\s--suite\s+(?P<suite>\S+))"