Skip to content

[https://nvbugs/5628952][fix] avoid cudaFree overlap with cuda graph #8903

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: release/1.1
Choose a base branch
from
+3 −4
Open

[https://nvbugs/5628952][fix] avoid cudaFree overlap with cuda graph #8903

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions tensorrt_llm/_torch/pyexecutor/py_executor_creator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
import copy
import enum
import gc
import importlib
import os
from concurrent.futures import ThreadPoolExecutor
Expand Down Expand Up @@ -600,6 +601,8 @@ def drafting_loop_wrapper(model):
with mem_monitor.observe_creation_stage(
_ExecutorCreationStage.INIT_EXTRA_RESOURCES
if estimating_kv_cache else _ExecutorCreationStage.EXTRA_RESOURCES):
# run gc.collect() to free memory of the previous py_executor, avoid cudaFree overlap with cuda graph capture
gc.collect()
py_executor = create_py_executor_instance(
dist=dist,
resources=resources,
Expand Down
4 changes: 0 additions & 4 deletions tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -345,10 +345,6 @@ full:H20-3e/test_e2e.py::test_ptp_quickstart_advanced_multi_gpus[DeepSeek-V3-671
disaggregated/test_disaggregated_single_gpu.py::test_disaggregated_spec_dec_batch_slot_limit[False-False-EAGLE3-LLaMA3.1-Instruct-8B-Llama-3.1-8B-Instruct] SKIP (https://nvbugs/5608743)
accuracy/test_cli_flow.py::TestMinitron4BBase::test_fp8 SKIP (https://nvbugs/5606233)
examples/test_gpt.py::test_llm_minitron_fp8_with_pseudo_loras[4b] SKIP (https://nvbugs/5606233)
disaggregated/test_disaggregated_single_gpu.py::test_disaggregated_simple_deepseek[True-False-DeepSeek-V3-Lite-fp8/fp8] SKIP (https://nvbugs/5626197)
disaggregated/test_disaggregated_single_gpu.py::test_disaggregated_simple_deepseek[True-True-DeepSeek-V3-Lite-fp8/fp8] SKIP (https://nvbugs/5628952)
cpp/test_e2e.py::test_benchmarks[t5-90] SKIP (https://nvbugs/5630196)
accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_nvfp4_4gpus[latency_moe_trtllm_eagle3] SKIP (https://nvbugs/5630700)
accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_nixl_backend SKIP (https://nvbugs/5628952)
accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_guided_decoding[xgrammar-mtp_nextn=2] SKIP (https://nvbugs/5628952)
full:RTX_PRO_6000_Blackwell_Server_Edition/accuracy/test_llm_api_pytorch.py::TestGemma3_1BInstruct::test_auto_dtype SKIP (https://nvbugs/5569696)