diff options
author | Hiroshi Yamauchi <yamauchi@google.com> | 2019-04-15 16:49:00 +0000 |
---|---|---|
committer | Hiroshi Yamauchi <yamauchi@google.com> | 2019-04-15 16:49:00 +0000 |
commit | 09e539fcaebb6362795d352cdcf4a818cf4d0d6a (patch) | |
tree | a91b7d07d85c67b737db8e49e3f62f95dd512898 /llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp | |
parent | 64041d7b90714aa6f0542c64ec5bb327beecfc58 (diff) | |
download | bcm5719-llvm-09e539fcaebb6362795d352cdcf4a818cf4d0d6a.tar.gz bcm5719-llvm-09e539fcaebb6362795d352cdcf4a818cf4d0d6a.zip |
[PGO] Profile guided code size optimization.
Summary:
Enable some of the existing size optimizations for cold code under PGO.
A ~5% code size saving in big internal app under PGO.
The way it gets BFI/PSI is discussed in the RFC thread
http://lists.llvm.org/pipermail/llvm-dev/2019-March/130894.html
Note it doesn't currently touch loop passes.
Reviewers: davidxl, eraman
Reviewed By: eraman
Subscribers: mgorny, javed.absar, smeenai, mehdi_amini, eraman, zzheng, steven_wu, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59514
llvm-svn: 358422
Diffstat (limited to 'llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp')
-rw-r--r-- | llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp index 2cc2e824cfb..d8b188ab8f8 100644 --- a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp +++ b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp @@ -4178,7 +4178,7 @@ Instruction *InstCombiner::tryOptimizeCall(CallInst *CI) { auto InstCombineErase = [this](Instruction *I) { eraseInstFromFunction(*I); }; - LibCallSimplifier Simplifier(DL, &TLI, ORE, InstCombineRAUW, + LibCallSimplifier Simplifier(DL, &TLI, ORE, BFI, PSI, InstCombineRAUW, InstCombineErase); if (Value *With = Simplifier.optimizeCall(CI)) { ++NumSimplified; |