diff options
| author | Connor Abbott <cwabbott0@gmail.com> | 2017-08-04 18:36:54 +0000 |
|---|---|---|
| committer | Connor Abbott <cwabbott0@gmail.com> | 2017-08-04 18:36:54 +0000 |
| commit | 66b9bd6e503bdd87616ab04509ffb3f70220c281 (patch) | |
| tree | ef831c300d054fcfca5e13142fa969db41c3fbf5 /llvm/test/Object/check_binary_output.ll | |
| parent | 92638ab62559ce082c02680efca7280344ca05dd (diff) | |
| download | bcm5719-llvm-66b9bd6e503bdd87616ab04509ffb3f70220c281.tar.gz bcm5719-llvm-66b9bd6e503bdd87616ab04509ffb3f70220c281.zip | |
[AMDGPU] Implement llvm.amdgcn.set.inactive intrinsic
Summary:
This intrinsic lets us set inactive lanes to an identity value when
implementing wavefront reductions. In combination with Whole Wavefront
Mode, it lets inactive lanes be skipped over as required by GLSL/Vulkan.
Lowering the intrinsic needs to happen post-RA so that RA knows that the
destination isn't completely overwritten due to the EXEC shenanigans, so
we need another pseudo-instruction to represent the un-lowered
intrinsic.
Reviewers: tstellar, arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye
Differential Revision: https://reviews.llvm.org/D34719
llvm-svn: 310088
Diffstat (limited to 'llvm/test/Object/check_binary_output.ll')
0 files changed, 0 insertions, 0 deletions
