BBVectorize: Better use of TTI->getShuffleCost

When flipping the pair of subvectors that form a vector, if the
vector length is 2, we can use the SK_Reverse shuffle kind to get
more-accurate cost information. Also we can use the SK_ExtractSubvector
shuffle kind to get accurate subvector extraction costs.

The current cost model implementations don't yet seem complex enough
for this to make a difference (thus, there are no test cases with this
commit), but it should help in future.

Depending on how the various targets optimize and combine shuffles in
practice, we might be able to get more-accurate costs by combining the
costs of multiple shuffle kinds. For example, the cost of flipping the
subvector pairs could be modeled as two extractions and two subvector
insertions. These changes, however, should probably be motivated
by specific test cases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173621 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 23 insertions, 4 deletions

diff --git a/lib/Transforms/Vectorize/BBVectorize.cpp b/lib/Transforms/Vectorize/BBVectorize.cpp
index 4bd11c6..9bf09ff 100644
--- a/lib/Transforms/Vectorize/BBVectorize.cpp
+++ b/lib/Transforms/Vectorize/BBVectorize.cpp
@@ -1824,6 +1824,16 @@ namespace {
                   R->second == PairConnectionSplat) {
                 int ESContrib = (int) getInstrCost(Instruction::ShuffleVector,
                                                    VTy, VTy);
+
+                if (VTy->getVectorNumElements() == 2) {
+                  if (R->second == PairConnectionSplat)
+                    ESContrib = std::min(ESContrib, (int) TTI->getShuffleCost(
+                      TargetTransformInfo::SK_Broadcast, VTy));
+                  else
+                    ESContrib = std::min(ESContrib, (int) TTI->getShuffleCost(
+                      TargetTransformInfo::SK_Reverse, VTy));
+                }
+
                 DEBUG(if (DebugPairSelection) dbgs() << "\tcost {" <<
                   *Q->second.first << " <-> " << *Q->second.second <<
                     "} -> {" <<
@@ -1860,10 +1870,12 @@ namespace {
 
             if (NeedsExtraction) {
               int ESContrib;
-              if (Ty1->isVectorTy())
+              if (Ty1->isVectorTy()) {
                 ESContrib = (int) getInstrCost(Instruction::ShuffleVector,
                                                Ty1, VTy);
-              else
+                ESContrib = std::min(ESContrib, (int) TTI->getShuffleCost(
+                  TargetTransformInfo::SK_ExtractSubvector, VTy, 0, Ty1));
+              } else
                 ESContrib = (int) TTI->getVectorInstrCost(
                                     Instruction::ExtractElement, VTy, 0);
 
@@ -1890,10 +1902,13 @@ namespace {
 
             if (NeedsExtraction) {
               int ESContrib;
-              if (Ty2->isVectorTy())
+              if (Ty2->isVectorTy()) {
                 ESContrib = (int) getInstrCost(Instruction::ShuffleVector,
                                                Ty2, VTy);
-              else
+                ESContrib = std::min(ESContrib, (int) TTI->getShuffleCost(
+                  TargetTransformInfo::SK_ExtractSubvector, VTy,
+                  Ty1->isVectorTy() ? Ty1->getVectorNumElements() : 1, Ty2));
+              } else
                 ESContrib = (int) TTI->getVectorInstrCost(
                                     Instruction::ExtractElement, VTy, 1);
               DEBUG(if (DebugPairSelection) dbgs() << "\tcost {" <<
@@ -1969,6 +1984,10 @@ namespace {
               } else if (IncomingPairs.count(VPR)) {
                 ESContrib = (int) getInstrCost(Instruction::ShuffleVector,
                                                VTy, VTy);
+
+                if (VTy->getVectorNumElements() == 2)
+                  ESContrib = std::min(ESContrib, (int) TTI->getShuffleCost(
+                    TargetTransformInfo::SK_Reverse, VTy));
               } else if (!Ty1->isVectorTy() && !Ty2->isVectorTy()) {
                 ESContrib = (int) TTI->getVectorInstrCost(
                                     Instruction::InsertElement, VTy, 0);