unset torch arch list for JIT mode #1765
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
jeffra
requested review from
samyam,
tjruwase,
ShadenSmith,
conglongli,
awan-10,
cli99,
eltonzheng,
minjiaz and
RezaYazdaniAminabadi
as code owners
RezaYazdaniAminabadi
approved these changes
Feb 11, 2022
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Outdated suggestions cannot be applied.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.
PyTorch will add
-gencodeargs for whatever CCs are in theTORCH_CUDA_ARCH_LISTat compile time, this makes sense when pre-compiling but does not make sense in JIT compile mode since we know at runtime what the targets need to be. Our op-builder has already set the intended targets that are needed at runtime. We need to unset the arch list before JIT compiling to avoid CC collisions such as multiple __half implementations. This PR simply stashes the arch list before JIT compiling and resets it after build.