[llvm] r318680 - [SelectionDAG] Make sorting predicate stronger to remove non-deterministic ordering
Mandeep Singh Grang via llvm-commits
llvm-commits at lists.llvm.org
Mon Nov 20 10:46:11 PST 2017
Author: mgrang
Date: Mon Nov 20 10:46:11 2017
New Revision: 318680
URL: http://llvm.org/viewvc/llvm-project?rev=318680&view=rev
Log:
[SelectionDAG] Make sorting predicate stronger to remove non-deterministic ordering
Summary:
This fixes failures in the following tests uncovered by D39245:
LLVM :: CodeGen/ARM/ifcvt3.ll
LLVM :: CodeGen/ARM/switch-minsize.ll
LLVM :: CodeGen/X86/switch.ll
Reviewers: hans, efriedma
Reviewed By: hans
Subscribers: fhahn, aemerson, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D39995
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
Modified: llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp?rev=318680&r1=318679&r2=318680&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp Mon Nov 20 10:46:11 2017
@@ -9350,10 +9350,12 @@ bool SelectionDAGBuilder::buildBitTests(
BitTestInfo BTI;
std::sort(CBV.begin(), CBV.end(), [](const CaseBits &a, const CaseBits &b) {
- // Sort by probability first, number of bits second.
+ // Sort by probability first, number of bits second, bit mask third.
if (a.ExtraProb != b.ExtraProb)
return a.ExtraProb > b.ExtraProb;
- return a.Bits > b.Bits;
+ if (a.Bits != b.Bits)
+ return a.Bits > b.Bits;
+ return a.Mask < b.Mask;
});
for (auto &CB : CBV) {
@@ -9542,10 +9544,13 @@ void SelectionDAGBuilder::lowerWorkItem(
}
if (TM.getOptLevel() != CodeGenOpt::None) {
- // Order cases by probability so the most likely case will be checked first.
+ // Here, we order cases by probability so the most likely case will be
+ // checked first. However, two clusters can have the same probability in
+ // which case their relative ordering is non-deterministic. So we use Low
+ // as a tie-breaker as clusters are guaranteed to never overlap.
std::sort(W.FirstCluster, W.LastCluster + 1,
[](const CaseCluster &a, const CaseCluster &b) {
- return a.Prob > b.Prob;
+ return a.Prob != b.Prob ? a.Prob > b.Prob : a.Low < b.Low;
});
// Rearrange the case blocks so that the last one falls through if possible
More information about the llvm-commits
mailing list