fix multi-thread exec of trt, test=release/1.5 #19379
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Shixiaowei02
force-pushed
the
1.5_fix/trt
branch
from
a9ae9a5 to
d68a71c
Compare
Shixiaowei02
force-pushed
the
1.5_fix/trt
branch
from
ac7cda3 to
9cd9d84
Compare
lijianshe02
approved these changes
Aug 28, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
[Cherry-pick of commit e1eca83, test=release/1.5]
The problem is only triggered by TensorRT and Clone(), and the user feedbacks that using version 1.4 can avoid this problem. There are two reasons for the final confirmation:
TensorRT uses the uniqueness of the molecular graph of the predictor_id area. Predictor::Clone() will change the predictor_id value, but the logic of TensorRT subgraph passes the predictor_id by value when Pass, and the Clone() does not re-trigger Pass, so the uniqueness fails.
Prior to commit ae576f3, the Deserialize() call to the TensorRT subgraph occurred at tensorrt_engine_op.h, once for each cloned-predictor. After this commit, the call occurred at tensorrt_subgraph_pass.cc, only once in ir_pass of master-predictor.
This pull request fixed the bug.