[X86][SSE] Add cost model for BSWAP of vectors
The BSWAP of vector types is quite efficiently implemented using vector shuffles on SSE/AVX targets, we should reflect the typical cost of this to encourage vectorization.
Differential Revision: http://reviews.llvm.org/D21521
llvm-svn: 273217
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index b5de8f9..338472a 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -951,19 +951,33 @@
{ ISD::BITREVERSE, MVT::v4i64, 5 },
{ ISD::BITREVERSE, MVT::v8i32, 5 },
{ ISD::BITREVERSE, MVT::v16i16, 5 },
- { ISD::BITREVERSE, MVT::v32i8, 5 }
+ { ISD::BITREVERSE, MVT::v32i8, 5 },
+ { ISD::BSWAP, MVT::v4i64, 1 },
+ { ISD::BSWAP, MVT::v8i32, 1 },
+ { ISD::BSWAP, MVT::v16i16, 1 }
};
static const CostTblEntry AVX1CostTbl[] = {
{ ISD::BITREVERSE, MVT::v4i64, 10 },
{ ISD::BITREVERSE, MVT::v8i32, 10 },
{ ISD::BITREVERSE, MVT::v16i16, 10 },
- { ISD::BITREVERSE, MVT::v32i8, 10 }
+ { ISD::BITREVERSE, MVT::v32i8, 10 },
+ { ISD::BSWAP, MVT::v4i64, 4 },
+ { ISD::BSWAP, MVT::v8i32, 4 },
+ { ISD::BSWAP, MVT::v16i16, 4 }
};
static const CostTblEntry SSSE3CostTbl[] = {
{ ISD::BITREVERSE, MVT::v2i64, 5 },
{ ISD::BITREVERSE, MVT::v4i32, 5 },
{ ISD::BITREVERSE, MVT::v8i16, 5 },
- { ISD::BITREVERSE, MVT::v16i8, 5 }
+ { ISD::BITREVERSE, MVT::v16i8, 5 },
+ { ISD::BSWAP, MVT::v2i64, 1 },
+ { ISD::BSWAP, MVT::v4i32, 1 },
+ { ISD::BSWAP, MVT::v8i16, 1 }
+ };
+ static const CostTblEntry SSE2CostTbl[] = {
+ { ISD::BSWAP, MVT::v2i64, 7 },
+ { ISD::BSWAP, MVT::v4i32, 7 },
+ { ISD::BSWAP, MVT::v8i16, 7 }
};
unsigned ISD = ISD::DELETED_NODE;
@@ -973,6 +987,9 @@
case Intrinsic::bitreverse:
ISD = ISD::BITREVERSE;
break;
+ case Intrinsic::bswap:
+ ISD = ISD::BSWAP;
+ break;
}
// Legalize the type.
@@ -996,6 +1013,10 @@
if (const auto *Entry = CostTableLookup(SSSE3CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
+ if (ST->hasSSE2())
+ if (const auto *Entry = CostTableLookup(SSE2CostTbl, ISD, MTy))
+ return LT.first * Entry->Cost;
+
return BaseT::getIntrinsicInstrCost(IID, RetTy, Tys, FMF);
}