kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-13 22:18:36 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[https://nvbugs/5643787][fix] remove the war path for notify to itself (#9834)

Browse Source 
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
This commit is contained in:
Chuang Zhu 2025-12-13 00:10:05 +08:00 committed by GitHub
parent 2fec53dfa5
commit 9c59c9f920
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 4 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
cpp/tensorrt_llm/executor/cache_transmission/nixl_utils/transferAgent.cpp
View File

@ -454,21 +454,10 @@ void NixlTransferAgent::invalidateRemoteAgent(std::string const& name)
void NixlTransferAgent::notifySyncMessage(std::string const& name, SyncMessage const& syncMessage)
{
if (name == mName)
{
// FIXME: nixl does not support gen notif to itself ,but support local transfer. we use local transfer to notify
// itself
MemoryDescs descs{MemoryType::kDRAM, {MemoryDesc{mDRamSrcBuffer}, MemoryDesc{mDRamDstBuffer}}};
TransferRequest request{TransferOp::kWRITE, descs, descs, name, syncMessage};
auto request_status = submitTransferRequests(request);
request_status->wait();
}
else
{
auto status = mRawAgent->genNotif(name, syncMessage);
TLLM_CHECK_WITH_INFO(
status == NIXL_SUCCESS, "genNotif failed with status: %s", nixlEnumStrings::statusStr(status).c_str());
}
auto status = mRawAgent->genNotif(name, syncMessage);
TLLM_CHECK_WITH_INFO(
status == NIXL_SUCCESS, "genNotif failed with status: %s", nixlEnumStrings::statusStr(status).c_str());
}
[[nodiscard]] std::unordered_map<std::string, std::vector<SyncMessage>> NixlTransferAgent::getNotifiedSyncMessages()

1
tests/integration/test_lists/waives.txt
View File

@ -411,7 +411,6 @@ accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_fp8_blockscale[throughput
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_nvfp4_4gpus[moe_backend=TRTLLM-mtp_nextn=0-ep4-fp8kv=True-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=False] SKIP (https://nvbugs/5721661)
accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput_mtp] SKIP (https://nvbugs/5715568)
unittest/_torch/modules/test_fused_moe.py::test_fused_moe_w4a8_nvfp4_fp8[CUTLASS] SKIP (https://nvbugs/5721912)
disaggregated/test_disaggregated.py::test_disaggregated_mixed[TinyLlama-1.1B-Chat-v1.0] SKIP (https://nvbugs/5661926)
unittest/_torch/auto_deploy/unit/singlegpu/custom_ops/test_flashinfer_attention_op.py::test_flashinfer_attention_op_context_input_pos[cuda-dtype0-4-8-seq6] SKIP (https://nvbugs/5721907)
disaggregated/test_auto_scaling.py::test_worker_restart[etcd-load_balancing] SKIP (https://nvbugs/5726066)
disaggregated/test_auto_scaling.py::test_worker_restart[http-kv_cache_aware] SKIP (https://nvbugs/5726066)