Tweak the ThinLTO pass pipeline

[mehdi_amini]

Summary

The original ThinLTO pipeline was derived from some
work I did tuning FullLTO on the test suite and SPEC. This
patch reduces the amount of work done in the "linker phase" of
the build, and extend the function simplifications passes
performed during the "compile phase". This helps the build time
by reducing the IR as much as possible during the compile phase
and limiting the work to be performed during the "link phase",
while keeping the performance "on par" with the existing pipeline.

[mehdi_amini]

Fix typo check for LTO

[mehdi_amini]

Rebase after splitting other changes

[tejohnson]

I compared performance of all the C/C++ benchmarks in SPEC cpu2006 on an Ivybridge X86 with gold+ThinLTO, with 3 runs each. It looks performance neutral (some run-to-run variations that look like noise, nothing significant).

lib/Transforms/IPO/PassManagerBuilder.cpp

428

Is there any significance to the move of the globalopt invocation here from its original location below LoopVersioningLICM?

448

Why are the globalDCE before and after globalopt no longer needed (for the reasons stated in the comments)?

[mehdi_amini]

lib/Transforms/IPO/PassManagerBuilder.cpp

428

I didn't really design the original position to be after LICM. I wrote this back in October before ReversePostOrderFunctionAttrr and LoopVersioningLICM exist.
I inserted it originally right after EliminateAvailableExternally.
In the meantime LoopVersioningLICM was inserted before EliminateAvailableExternally. In this patch I reorganized ReversePostOrderFunctionAttrr, EliminateAvailableExternally, and GlobalOpt to be ran before LICM (which is more part of the function optimization pipeline that follow) last week in this patch.
Then I was asked to split the patch, so what remains is only this part of the move.

448

I think my original comments were buggy: GlobalOptimizer performs DCE as well.

[mehdi_amini]

r268769

lib/Transforms/IPO/PassManagerBuilder.cpp

void PassManagerBuilder::addFunctionSimplificationPasses(
   MPM.add(createCFGSimplificationPass());     // Merge & remove BBs
   // Combine silly seq's
   addInstructionCombiningPass(MPM);
   addExtensionsToPM(EP_Peephole, MPM);
 
   MPM.add(createTailCallEliminationPass()); // Eliminate tail calls
   MPM.add(createCFGSimplificationPass());     // Merge & remove BBs
   MPM.add(createReassociatePass());           // Reassociate expressions
-  if (PrepareForThinLTO) {
-    MPM.add(createAggressiveDCEPass());        // Delete dead instructions
-    addInstructionCombiningPass(MPM);          // Combine silly seq's
-    return;
-  }
   // Rotate Loop - disable header duplication at -Oz
   MPM.add(createLoopRotatePass(SizeLevel == 2 ? 0 : -1));
   MPM.add(createLICMPass());                  // Hoist loop invariants
   MPM.add(createLoopUnswitchPass(SizeLevel || OptLevel < 3));
   MPM.add(createCFGSimplificationPass());
   addInstructionCombiningPass(MPM);
   MPM.add(createIndVarSimplifyPass());        // Canonicalize indvars
   MPM.add(createLoopIdiomPass());             // Recognize idioms like memset.
void PassManagerBuilder::populateModulePassManager(
   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;
   }
 
+  if (PerformThinLTO)
+    // Optimize globals now when performing ThinLTO, this enables more
+    // optimizations later.
+    MPM.add(createGlobalOptimizerPass());
+
   // Scheduling LoopVersioningLICM when inlining is over, because after that
   // we may see more accurate aliasing. Reason to run this late is that too
   // early versioning may prevent further inlining due to increase of code
-  // size. By placing it just after inlining other optimizations which runs 
-  // later might get benefit of no-alias assumption in clone loop. 
+  // size. By placing it just after inlining other optimizations which runs
+  // later might get benefit of no-alias assumption in clone loop.
   if (UseLoopVersioningLICM) {
     MPM.add(createLoopVersioningLICMPass());    // Do LoopVersioningLICM
     MPM.add(createLICMPass());                  // Hoist loop invariants
   }
 
-  if (PerformThinLTO) {
-    // Remove dead fns and globals. Removing unreferenced functions could lead
-    // to more opportunities for globalopt.
-    MPM.add(createGlobalDCEPass());
-    MPM.add(createGlobalOptimizerPass());
-    // Remove dead fns and globals after globalopt.
-    MPM.add(createGlobalDCEPass());
-    addFunctionSimplificationPasses(MPM);
-  }
-
   if (EnableNonLTOGlobalsModRef)
     // We add a fresh GlobalsModRef run at this point. This is particularly
     // useful as the above will have inlined, DCE'ed, and function-attr
     // propagated everything. We should at this point have a reasonably minimal
     // and richly annotated call graph. By computing aliasing and mod/ref
     // information for all local globals here, the late loop passes and notably
     // the vectorizer will be able to use them to help recognize vectorizable
     // memory operations.