Fix compilation with clang on ARM64 #1285
Open
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Outdated suggestions cannot be applied.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.
While compiling with clang-17 on ARM64
It seems that in pytorch/pytorch@10f3abc (first released in 2.3.0)
get_device()
was changed to return ac10::DeviceIndex
. Sincec10::DeviceIndex
is an alias toint8_t
andchar
is unsigned on ARM, this is a narrowing conversion.Simply removing the cast entirely would work on pytorch >= 2.3.0, but break on older versions.
Instead, switch to the
device()
API for which there is already aCUDAGuard
constructor.