diff options
| author | Kazu Hirata <kazu@google.com> | 2020-06-04 00:40:17 -0700 |
|---|---|---|
| committer | Kazu Hirata <kazu@google.com> | 2020-06-04 00:40:58 -0700 |
| commit | 347a599e5f03ab708943559d0a131127b206b576 (patch) | |
| tree | db3111702ba3f4c83f455277ae78259971928801 /llvm/lib/Analysis/InlineCost.cpp | |
| parent | 7eff1a7136049fd38d779f52f905f88181b03679 (diff) | |
| download | llvm-347a599e5f03ab708943559d0a131127b206b576.zip llvm-347a599e5f03ab708943559d0a131127b206b576.tar.gz llvm-347a599e5f03ab708943559d0a131127b206b576.tar.bz2 | |
[Inlining] Introduce -enable-npm-pgo-inline-deferral
Summary:
Experiments show that inline deferral past pre-inlining slightly
pessimizes the performance.
This patch introduces an option to control inline deferral during PGO.
The option defaults to true for now (that is, NFC).
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: eraman, hiraditya, haicheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80776
Diffstat (limited to 'llvm/lib/Analysis/InlineCost.cpp')
| -rw-r--r-- | llvm/lib/Analysis/InlineCost.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/llvm/lib/Analysis/InlineCost.cpp b/llvm/lib/Analysis/InlineCost.cpp index 7377b088..9eea2e9 100644 --- a/llvm/lib/Analysis/InlineCost.cpp +++ b/llvm/lib/Analysis/InlineCost.cpp @@ -2250,7 +2250,8 @@ Optional<int> llvm::getInliningCostEstimate( /*HotCallSiteThreshold*/ {}, /*LocallyHotCallSiteThreshold*/ {}, /*ColdCallSiteThreshold*/ {}, - /* ComputeFullInlineCost*/ true}; + /*ComputeFullInlineCost*/ true, + /*EnableDeferral*/ true}; InlineCostCallAnalyzer CA(*Call.getCalledFunction(), Call, Params, CalleeTTI, GetAssumptionCache, GetBFI, PSI, ORE, true, |