[CSSPGO][Preinliner] Bump up the threshold to favor previous compiler inline decision.

[hoy]

Summary

The compiler has more insight and knowledge about functions based on their IR and attribures and should make a better inline decision than the offline preinliner does which is purely based on callsites hotness and code size. Therefore I'm making changes to favor previous compiler inline decision by bumping up the callsite allowance.

This should improve the performance by more than 1% according to testing on Meta services.

[hoy]

Updating D153797: [CSSPGO][Preinliner] Bump up the threshold to favor previous compiler inline decision.

[wenlei]

llvm/tools/llvm-profgen/CSPreInliner.cpp

199

Make this a tunable value via a switch?

[hoy]

llvm/tools/llvm-profgen/CSPreInliner.cpp

199

Sounds good.

Forgot to mention that smaller numbers like 10, 25, 50 don't give as good win as 100 gives.

[hoy]

Addressing feedbacks.

[wenlei]

lgtm, thanks.

[wenlei]

llvm/tools/llvm-profgen/CSPreInliner.cpp

61

multipifier->multiplier for all instances in switch name, variable names.

61

and switch desc too.

[hoy]

llvm/tools/llvm-profgen/CSPreInliner.cpp

61

Good catch.

[hoy]

Updating D153797: [CSSPGO][Preinliner] Bump up the threshold to favor previous compiler inline decision.

llvm/tools/llvm-profgen/CSPreInliner.cpp

cl::opt<bool> UseContextCostForPreInliner(
     cl::desc("Use context-sensitive byte size cost for preinliner decisions"));
 } // namespace llvm
 
 static cl::opt<bool> SamplePreInlineReplay(
     "csspgo-replay-preinline", cl::Hidden, cl::init(false),
     cl::desc(
         "Replay previous inlining and adjust context profile accordingly"));
 
+static cl::opt<int> CSPreinlMultiplierForPrevInl(
+    "csspgo-preinliner-multiplier-for-previous-inlining", cl::Hidden,
+    cl::init(100),
+    cl::desc(
+        "Multiplier to bump up callsite threshold for previous inlining."));
+
 CSPreInliner::CSPreInliner(SampleContextTracker &Tracker,
                            ProfiledBinary &Binary, ProfileSummary *Summary)
     : UseContextCost(UseContextCostForPreInliner),
       // TODO: Pass in a guid-to-name map in order for
       // ContextTracker.getFuncNameFor to work, if `Profiles` can have md5 codes
       // as their profile context.
       ContextTracker(Tracker), Binary(Binary), Summary(Summary) {
   // Set default preinliner hot/cold call site threshold tuned with CSSPGO.

 uint32_t CSPreInliner::getFuncSize(const ContextTrieNode *ContextNode) {
   if (UseContextCost)
     return Binary.getFuncSizeForContext(ContextNode);
 
   return ContextNode->getFunctionSamples()->getBodySamples().size();
 }
 
 bool CSPreInliner::shouldInline(ProfiledInlineCandidate &Candidate) {
+  bool WasInlined =
+      Candidate.CalleeSamples->getContext().hasAttribute(ContextWasInlined);
   // If replay inline is requested, simply follow the inline decision of the
   // profiled binary.
   if (SamplePreInlineReplay)
-    return Candidate.CalleeSamples->getContext().hasAttribute(
-        ContextWasInlined);
+    return WasInlined;
 
   unsigned int SampleThreshold = SampleColdCallSiteThreshold;
   uint64_t ColdCountThreshold = ProfileSummaryBuilder::getColdCountThreshold(
       (Summary->getDetailedSummary()));
 
   if (Candidate.CallsiteCount <= ColdCountThreshold)
     SampleThreshold = SampleColdCallSiteThreshold;
   else {
double NormalizedHotness =
         (NormalizationUpperBound - NormalizationLowerBound);
     if (NormalizedHotness > 1.0)
       NormalizedHotness = 1.0;
     // Add 1 to to ensure hot callsites get a non-zero threshold, which could
     // happen when SampleColdCallSiteThreshold is 0. This is when we do not
     // want any inlining for cold callsites.
     SampleThreshold = SampleHotCallSiteThreshold * NormalizedHotness * 100 +
                       SampleColdCallSiteThreshold + 1;
+    // Bump up the threshold to favor previous compiler inline decision. The
+    // compiler has more insight and knowledge about functions based on their IR
+    // and attribures and should be able to make a more reasonable inline
+    // decision.
+    if (WasInlined)
+      SampleThreshold *= CSPreinlMultiplierForPrevInl;
   }
 
   return (Candidate.SizeCost < SampleThreshold);
 }
 
 void CSPreInliner::processFunction(const StringRef Name) {
   FunctionSamples *FSamples = ContextTracker.getBaseSamplesFor(Name);
   if (!FSamples)