Run GlobalOpt before emitting the bitcode for ThinLTO

This is motivated by reducing the size of the IR and thus reduce
compile time.

From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267385

1 files changed, 2 insertions, 0 deletions

diff --git a/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp b/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
index 7867bf44a52..0dc44c660f1 100644
--- a/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
+++ b/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
@@ -404,6 +404,8 @@ void PassManagerBuilder::populateModulePassManager(
   // unrolling/vectorization/... now. We'll first run the inliner + CGSCC passes
   // during ThinLTO and perform the rest of the optimizations afterward.
   if (PrepareForThinLTO) {
+    // Reduce the size of the IR as much as possible.
+    MPM.add(createGlobalOptimizerPass());
     // Rename anon function to be able to export them in the summary.
     MPM.add(createNameAnonFunctionPass());
     return;