Page MenuHomePhabricator

[AMDGPU] Enable the implicit arguments for HIP (CLANG)
ClosedPublic

Authored by cdevadas on May 22 2019, 4:08 AM.

Details

Summary

Enable 48-bytes of implicit arguments for HIP as well. Earlier it was enabled for OpenCL. This code is specific to AMDGPU target.

Diff Detail

Repository
rL LLVM

Event Timeline

cdevadas created this revision.May 22 2019, 4:08 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 22 2019, 4:08 AM

Currently HIP and CUDA share the same test directories, so better put the test in CodeGenCUDA.

cdevadas updated this revision to Diff 200870.May 22 2019, 10:29 PM

Moved the test to CodeGenCUDA directory.

hliao added a subscriber: hliao.Mon, Jun 10, 9:09 AM
hliao added inline comments.
test/CodeGenCUDA/amdgpu-hip-implicit-kernarg.cu
7 ↗(On Diff #200870)

For a single check, you don't need CHECK-DAG.

cdevadas updated this revision to Diff 203975.Mon, Jun 10, 10:29 PM

simplified the check in the test case.

yaxunl accepted this revision.Thu, Jun 13, 12:15 PM
This revision is now accepted and ready to land.Thu, Jun 13, 12:15 PM
This revision was automatically updated to reflect the committed changes.
Herald added a project: Restricted Project. · View Herald TranscriptFri, Jun 14, 8:52 AM