[SystemZ::TTI] Improve accuracy of costs for vector fp <-> int conversions

Improve getCastInstrCost() by respecting the different types of Src and Dst
for vector integer <-> fp conversions.

This means that extracting from integer becomes more expensive (by the
extraction penalty), and the extraction from fp becomes cheaper (no longer
has a false extraction penalty).

Review: Ulrich Weigand
https://reviews.llvm.org/D54423

llvm-svn: 346663

1 files changed, 2 insertions, 1 deletions

diff --git a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
index f296d80dbf5..b25059eb3c0 100644
--- a/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
+++ b/llvm/lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
@@ -718,7 +718,8 @@ int SystemZTTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
           (Opcode == Instruction::FPToSI || Opcode == Instruction::FPToUI))
         NeedsExtracts = false;
 
-      TotCost += getScalarizationOverhead(Dst, NeedsInserts, NeedsExtracts);
+      TotCost += getScalarizationOverhead(Src, false, NeedsExtracts);
+      TotCost += getScalarizationOverhead(Dst, NeedsInserts, false);
 
       // FIXME: VF 2 for float<->i32 is currently just as expensive as for VF 4.
       if (VF == 2 && SrcScalarBits == 32 && DstScalarBits == 32)