[XLA][AMDGPU] use AddrSpaceCast to access GlobalVariable on AMDGPU#20754
Merged
tensorflow-copybara merged 1 commit intotensorflow:masterfrom Jul 14, 2018
Conversation
In XLA kernels pointers are in default address space (0). On LLVM AMDGPU backend, all GlobalVariable instances must be in global address space (1). Therefore in an AddrSpaceCast is required for AMDGPU backend. To cope with existing NVPTX backend, use ConstantExpr::getPointerBitCastOrAddrSpaceCast in this commit.
Contributor
Author
|
@jlebar this is a patch in XLA to cope with the special property held by LLVM AMDGPU backend |
jlebar
approved these changes
Jul 12, 2018
Contributor
|
@qlzh727 can you help merge this? |
qlzh727
added
awaiting testing (then merge)
ready to pull
Contributor
Author
|
@qlzh727 the error in "Ubuntu Sanity" target seems to have nothing to do with this PR but a configuration in the test environment. is it possible to force restart the test? |
Contributor
|
I will attempt to merge this. |
tensorflow-copybara
pushed a commit
that referenced
this pull request
Jul 14, 2018
…xla-globalvariable-addrspacecast PiperOrigin-RevId: 204560820
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
In XLA kernels pointers are in default address space (0). On LLVM AMDGPU
backend, all GlobalVariable instances must be in global address space (1).
Therefore an AddrSpaceCast is required for AMDGPU backend. To cope with
existing NVPTX backend, use ConstantExpr::getPointerBitCastOrAddrSpaceCast
in this commit.