Cost Model: add tables for some avx type-conversion hacks.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167480 91177308-0d34-0410-b5e6-96231b3b80d8

4 files changed, 94 insertions, 3 deletions

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 6b5466d70b..5853b5109c 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -17526,6 +17526,23 @@ int FindInTable(const X86CostTblEntry *Tbl, unsigned len, int ISD, MVT Ty) {
   return -1;
 }
 
+struct X86TypeConversionCostTblEntry {
+  int ISD;
+  MVT Dst;
+  MVT Src;
+  unsigned Cost;
+};
+
+int FindInConvertTable(const X86TypeConversionCostTblEntry *Tbl, unsigned len,
+                       int ISD, MVT Dst, MVT Src) {
+  for (unsigned int i = 0; i < len; ++i)
+    if (Tbl[i].ISD == ISD && Tbl[i].Src == Src && Tbl[i].Dst == Dst)
+      return i;
+
+  // Could not find an entry.
+  return -1;
+}
+
 unsigned
 X86VectorTargetTransformInfo::getArithmeticInstrCost(unsigned Opcode,
                                                      Type *Ty) const {
@@ -17535,8 +17552,7 @@ X86VectorTargetTransformInfo::getArithmeticInstrCost(unsigned Opcode,
   int ISD = InstructionOpcodeToISD(Opcode);
   assert(ISD && "Invalid opcode");
 
-  const X86Subtarget &ST =
-  TLI->getTargetMachine().getSubtarget<X86Subtarget>();
+  const X86Subtarget &ST = TLI->getTargetMachine().getSubtarget<X86Subtarget>();
 
   static const X86CostTblEntry AVX1CostTable[] = {
     // We don't have to scalarize unsupported ops. We can issue two half-sized
@@ -17647,5 +17663,45 @@ unsigned X86VectorTargetTransformInfo::getCmpSelInstrCost(unsigned Opcode,
   return VectorTargetTransformImpl::getCmpSelInstrCost(Opcode, ValTy, CondTy);
 }
 
+unsigned X86VectorTargetTransformInfo::getCastInstrCost(unsigned Opcode,
+                                                        Type *Dst,
+                                                        Type *Src) const {
+  int ISD = InstructionOpcodeToISD(Opcode);
+  assert(ISD && "Invalid opcode");
+
+  EVT SrcTy = TLI->getValueType(Src);
+  EVT DstTy = TLI->getValueType(Dst);
+
+  if (!SrcTy.isSimple() || !DstTy.isSimple())
+    return VectorTargetTransformImpl::getCastInstrCost(Opcode, Dst, Src);
+
+  const X86Subtarget &ST = TLI->getTargetMachine().getSubtarget<X86Subtarget>();
+
+  static const X86TypeConversionCostTblEntry AVXConversionTbl[] = {
+    { ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i16, 1 },
+    { ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i16, 1 },
+    { ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 1 },
+    { ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 1 },
+    { ISD::TRUNCATE,    MVT::v4i32, MVT::v4i64, 1 },
+    { ISD::TRUNCATE,    MVT::v8i16, MVT::v8i32, 1 },
+    { ISD::SINT_TO_FP,  MVT::v8f32, MVT::v8i8,  1 },
+    { ISD::SINT_TO_FP,  MVT::v4f32, MVT::v4i8,  1 },
+    { ISD::UINT_TO_FP,  MVT::v8f32, MVT::v8i8,  1 },
+    { ISD::UINT_TO_FP,  MVT::v4f32, MVT::v4i8,  1 },
+    { ISD::FP_TO_SINT,  MVT::v8i8, MVT::v8f32,  1 },
+    { ISD::FP_TO_SINT,  MVT::v4i8, MVT::v4f32,  1 },
+    { ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i1,  6 },
+    { ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i1,  9 },
+  };
 
+  if (ST.hasAVX()) {
+    int Idx = FindInConvertTable(AVXConversionTbl,
+                                 array_lengthof(AVXConversionTbl),
+                                 ISD, DstTy.getSimpleVT(), SrcTy.getSimpleVT());
+    if (Idx != -1)
+      return AVXConversionTbl[Idx].Cost;
+  }
+
+  return VectorTargetTransformImpl::getCastInstrCost(Opcode, Dst, Src);
+}
 
diff --git a/lib/Target/X86/X86ISelLowering.h b/lib/Target/X86/X86ISelLowering.h
index 40dfaa0d43..c6354b0801 100644
--- a/lib/Target/X86/X86ISelLowering.h
+++ b/lib/Target/X86/X86ISelLowering.h
@@ -960,6 +960,9 @@ namespace llvm {
 
     unsigned getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
                                 Type *CondTy) const;
+
+    virtual unsigned getCastInstrCost(unsigned Opcode, Type *Dst,
+                                      Type *Src) const;
   };
 }
 
diff --git a/test/Analysis/CostModel/X86/cast.ll b/test/Analysis/CostModel/X86/cast.ll
index f8b1114a7c..298d5c66ad 100644
--- a/test/Analysis/CostModel/X86/cast.ll
+++ b/test/Analysis/CostModel/X86/cast.ll
@@ -32,3 +32,35 @@ define i32 @add(i32 %arg) {
   ret i32 undef
 }
 
+define i32 @zext_sext(<8 x i1> %in) {
+  ;CHECK: cost of 6 {{.*}} zext
+  %Z = zext <8 x i1> %in to <8 x i32>
+  ;CHECK: cost of 9 {{.*}} sext
+  %S = sext <8 x i1> %in to <8 x i32>
+
+  ;CHECK: cost of 1 {{.*}} sext
+  %A = sext <8 x i16> undef to <8 x i32>
+  ;CHECK: cost of 1 {{.*}} zext
+  %B = zext <8 x i16> undef to <8 x i32>
+  ;CHECK: cost of 1 {{.*}} sext
+  %C = sext <4 x i32> undef to <4 x i64>
+
+  ;CHECK: cost of 1 {{.*}} zext
+  %D = zext <4 x i32> undef to <4 x i64>
+  ;CHECK: cost of 1 {{.*}} trunc
+
+  %E = trunc <4 x i64> undef to <4 x i32>
+  ;CHECK: cost of 1 {{.*}} trunc
+  %F = trunc <8 x i32> undef to <8 x i16>
+
+  ret i32 undef
+}
+
+define i32 @masks(<8 x i1> %in) {
+  ;CHECK: cost of 6 {{.*}} zext
+  %Z = zext <8 x i1> %in to <8 x i32>
+  ;CHECK: cost of 9 {{.*}} sext
+  %S = sext <8 x i1> %in to <8 x i32>
+  ret i32 undef
+}
+
diff --git a/test/Transforms/LoopVectorize/X86/cost-model.ll b/test/Transforms/LoopVectorize/X86/cost-model.ll
index 40e660855b..628f9912c8 100644
--- a/test/Transforms/LoopVectorize/X86/cost-model.ll
+++ b/test/Transforms/LoopVectorize/X86/cost-model.ll
@@ -9,7 +9,7 @@ target triple = "x86_64-apple-macosx10.8.0"
 @a = common global [2048 x i32] zeroinitializer, align 16
 
 ;CHECK: cost_model_1
-;CHECK-NOT: <4 x i32>
+;CHECK: <4 x i32>
 ;CHECK: ret void
 define void @cost_model_1() nounwind uwtable noinline ssp {
 entry: