kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-02-04 18:21:52 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[TRTLLM-10415][feat] Dump thread stacks for hanging tests before time… (#10708)

Browse Source 
Signed-off-by: Fred Wei <20514172+WeiHaocheng@users.noreply.github.com>
This commit is contained in:
WeiHaocheng 2026-01-29 20:43:34 +08:00 committed by GitHub
parent c7a86f89de
commit 80dd6e70c6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
tests/unittest/conftest.py
View File

@ -14,6 +14,7 @@
# limitations under the License.
# # Force resource release after test
import os
import signal
import sys
import traceback
import warnings
@ -33,14 +34,23 @@ from mpi4py.futures import MPIPoolExecutor
from utils.cpp_paths import llm_root # noqa: F401
from utils.util import get_current_process_gpu_memory
from tensorrt_llm._utils import print_all_stacks
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from integration.defs import test_list_parser
def dump_threads(signum, frame):
print_all_stacks()
def pytest_configure(config):
# avoid thread leak of tqdm's TMonitor
tqdm.tqdm.monitor_interval = 0
# Dump all threads' stacks when SIGALRM is received
signal.signal(signal.SIGALRM, dump_threads)
@pytest.hookimpl(wrapper=True)
def pytest_runtest_protocol(item, nextitem):