Irreducible loop metadata for more accurate block frequency under PGO.
Summary:
Currently the block frequency analysis is an approximation for irreducible
loops.
The new irreducible loop metadata is used to annotate the irreducible loop
headers with their header weights based on the PGO profile (currently this is
approximated to be evenly weighted) and to help improve the accuracy of the
block frequency analysis for irreducible loops.
This patch is a basic support for this.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: mehdi_amini, llvm-commits, eraman
Differential Revision: https://reviews.llvm.org/D39028
llvm-svn: 317278
diff --git a/llvm/lib/Analysis/BlockFrequencyInfoImpl.cpp b/llvm/lib/Analysis/BlockFrequencyInfoImpl.cpp
index 1030407..7e32302 100644
--- a/llvm/lib/Analysis/BlockFrequencyInfoImpl.cpp
+++ b/llvm/lib/Analysis/BlockFrequencyInfoImpl.cpp
@@ -271,6 +271,7 @@
// Swap with a default-constructed std::vector, since std::vector<>::clear()
// does not actually clear heap storage.
std::vector<FrequencyData>().swap(Freqs);
+ IsIrrLoopHeader.clear();
std::vector<WorkingData>().swap(Working);
Loops.clear();
}
@@ -280,8 +281,10 @@
/// Releases all memory not used downstream. In particular, saves Freqs.
static void cleanup(BlockFrequencyInfoImplBase &BFI) {
std::vector<FrequencyData> SavedFreqs(std::move(BFI.Freqs));
+ SparseBitVector<> SavedIsIrrLoopHeader(std::move(BFI.IsIrrLoopHeader));
BFI.clear();
BFI.Freqs = std::move(SavedFreqs);
+ BFI.IsIrrLoopHeader = std::move(SavedIsIrrLoopHeader);
}
bool BlockFrequencyInfoImplBase::addToDist(Distribution &Dist,
@@ -572,6 +575,13 @@
return BlockCount.getLimitedValue();
}
+bool
+BlockFrequencyInfoImplBase::isIrrLoopHeader(const BlockNode &Node) {
+ if (!Node.isValid())
+ return false;
+ return IsIrrLoopHeader.test(Node.Index);
+}
+
Scaled64
BlockFrequencyInfoImplBase::getFloatingBlockFreq(const BlockNode &Node) const {
if (!Node.isValid())
@@ -819,3 +829,14 @@
DEBUG(debugAssign(*this, D, W.TargetNode, Taken, nullptr));
}
}
+
+void BlockFrequencyInfoImplBase::distributeIrrLoopHeaderMass(Distribution &Dist) {
+ BlockMass LoopMass = BlockMass::getFull();
+ DitheringDistributer D(Dist, LoopMass);
+ for (const Weight &W : Dist.Weights) {
+ BlockMass Taken = D.takeMass(W.Amount);
+ assert(W.Type == Weight::Local && "all weights should be local");
+ Working[W.TargetNode.Index].getMass() = Taken;
+ DEBUG(debugAssign(*this, D, W.TargetNode, Taken, nullptr));
+ }
+}