projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: f469c64) | patch
[AMDGPU] Prevent sequences of non-instructions disrupting GCNHazardRecognizer wait...
authorCarl Ritson <carl.ritson@amd.com>
Mon, 10 Sep 2018 10:14:48 +0000 (10:14 +0000)
committerCarl Ritson <carl.ritson@amd.com>
Mon, 10 Sep 2018 10:14:48 +0000 (10:14 +0000)
commitf898edd117daec42762c19dd733189307414d88a
treee394c9ef1bbd07b8de7dd5282d17cbaef9c5396atree | snapshot
parentf469c64fef0f13fe92a4578526c1f2f5161b81eccommit | diff
[AMDGPU] Prevent sequences of non-instructions disrupting GCNHazardRecognizer wait state counting

Summary:
This fixes a bug where a large number of implicit def instructions can fill the GCNHazardRecognizer lookahead buffer causing required NOPs to not be inserted.

Reviewers: nhaehnle, arsenm

Reviewed By: arsenm

Subscribers: sheredom, kzhuravl, jvesely, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D51726

Change-Id: Ie75338f94de704ee5816b05afd0c922c6748a95b
llvm-svn: 341798
llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/hazard.mir diff | blob | history
Domain: System / Toolchain;