Remove dead preprocessor code for number of CUDA threads #3285
Conversation
This file contains hidden or 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
This was referenced Nov 5, 2015
__CUDA_ARCH__ is not defined in host code; the #if was vacuous and misleading.
671c55b to
bc1aa41
Compare
shelhamer
added a commit
that referenced
this pull request
Dec 2, 2015
Remove dead preprocessor code for number of CUDA threads
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
See #3281, #418, and https://groups.google.com/forum/#!topic/caffe-users/4abF674UaYY/discussion.
The preprocessor check added by #62 is not correct. Although the code appears to check the CUDA architecture version, the
__CUDA_ARCH__macro is not defined in host code, soCAFFE_CUDA_NUM_THREADSalways gets set to 512.This patch simply removes the misleading dead code, and keeps
CAFFE_CUDA_NUM_THREADSat 512. Given that we've been using this value for 21 months, and to my knowledge there's no a priori compelling reason why it ought to be maxed out, I don't see any reason to add additional code right now to restore the intended behavior (although I'd welcome a future PR that does so, if it makes a compelling performance difference, which I don't expect it does).