From ee000f673458c4d78e5d802a2769990fdf5c303e Mon Sep 17 00:00:00 2001 From: Xiaoyu Zhang <1182563586@qq.com> Date: Tue, 14 Jul 2026 17:32:29 +0800 Subject: [PATCH] [CI] Fix SGLANG_JIT_KERNEL_RUN_FULL_TESTS never activating the nightly full jit-kernel sweep (#31042) Co-authored-by: Claude Fable 5 --- .github/workflows/nightly-test-nvidia.yml | 4 +++- python/sglang/jit_kernel/utils.py | 4 ++-- python/sglang/srt/environ.py | 2 ++ 3 files changed, 7 insertions(+), 3 deletions(-) diff --git a/.github/workflows/nightly-test-nvidia.yml b/.github/workflows/nightly-test-nvidia.yml index ed70fbbc7..a0714740c 100644 --- a/.github/workflows/nightly-test-nvidia.yml +++ b/.github/workflows/nightly-test-nvidia.yml @@ -145,7 +145,9 @@ jobs: timeout-minutes: 90 run: | cd test - python3 run_suite.py --hw cuda --suite nightly-kernel-8-gpu-h200 --nightly --continue-on-error + # Full grids run ~7x the in-CI parametrizations per world size; the + # default 1200s per-file budget only fits the reduced PR sweep. + python3 run_suite.py --hw cuda --suite nightly-kernel-8-gpu-h200 --nightly --continue-on-error --timeout-per-file 3600 - uses: ./.github/actions/upload-cuda-coredumps if: failure() diff --git a/python/sglang/jit_kernel/utils.py b/python/sglang/jit_kernel/utils.py index bdff4bb67..2ad5c25c4 100644 --- a/python/sglang/jit_kernel/utils.py +++ b/python/sglang/jit_kernel/utils.py @@ -24,19 +24,19 @@ from typing import ( import torch +from sglang.srt.environ import envs from sglang.utils import is_in_ci if TYPE_CHECKING: from tvm_ffi import Module F = TypeVar("F", bound=Callable[..., Any]) -_FULL_TEST_ENV_VAR = "SGLANG_JIT_KERNEL_RUN_FULL_TESTS" logger = logging.getLogger(__name__) def should_run_full_tests() -> bool: - return os.getenv(_FULL_TEST_ENV_VAR, "false").lower() == "true" + return envs.SGLANG_JIT_KERNEL_RUN_FULL_TESTS.get() def get_ci_test_range(full_range: List[Any], ci_range: List[Any]) -> List[Any]: diff --git a/python/sglang/srt/environ.py b/python/sglang/srt/environ.py index 910e74aa9..9de2fa48c 100644 --- a/python/sglang/srt/environ.py +++ b/python/sglang/srt/environ.py @@ -245,6 +245,8 @@ class Envs: # else /tmp); see debug_utils/cuda_coredump.py. SGLANG_CUDA_COREDUMP_DIR = EnvStr(None) SGLANG_TEST_MAX_RETRY = EnvInt(None) + # Expand jit_kernel test grids to their full parameter ranges (nightly). + SGLANG_JIT_KERNEL_RUN_FULL_TESTS = EnvBool(False) # Constrained Decoding (Grammar) SGLANG_GRAMMAR_POLL_INTERVAL = EnvFloat(0.005)