[CostModel][X86] Fix SLM <2 x i64> icmp costs

SLM is 2 x slower for <2 x i64> comparison ops than other vector types, we should account for this like we do for SLM <2 x i64> add/sub/mul costs.

This should remove some of the SLM codegen diffs in D43582

llvm-svn: 372954

1 files changed, 9 insertions, 0 deletions

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 10ece8a87c6..838a43ac339 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1797,6 +1797,11 @@ int X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy, Type *CondTy,
     }
   }
 
+  static const CostTblEntry SLMCostTbl[] = {
+    // slm pcmpeq/pcmpgt throughput is 2
+    { ISD::SETCC,   MVT::v2i64,   2 },
+  };
+
   static const CostTblEntry AVX512BWCostTbl[] = {
     { ISD::SETCC,   MVT::v32i16,  1 },
     { ISD::SETCC,   MVT::v64i8,   1 },
@@ -1883,6 +1888,10 @@ int X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy, Type *CondTy,
     { ISD::SELECT,  MVT::v4f32,   3 }, // andps + andnps + orps
   };
 
+  if (ST->isSLM())
+    if (const auto *Entry = CostTableLookup(SLMCostTbl, ISD, MTy))
+      return LT.first * (ExtraCost + Entry->Cost);
+
   if (ST->hasBWI())
     if (const auto *Entry = CostTableLookup(AVX512BWCostTbl, ISD, MTy))
       return LT.first * (ExtraCost + Entry->Cost);