optimize bitcast(trunc(bitcast(x))) where the result is a float and 'x'

is a vector to be a vector element extraction.  This allows clang to
compile:

struct S { float A, B, C, D; };
float foo(struct S A) { return A.A + A.B+A.C+A.D; }

into:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	movd	%xmm0, %rax
	shrq	$32, %rax
	movd	%eax, %xmm2
	addss	%xmm0, %xmm2
	movapd	%xmm1, %xmm3
	addss	%xmm2, %xmm3
	movd	%xmm1, %rax
	shrq	$32, %rax
	movd	%eax, %xmm0
	addss	%xmm3, %xmm0
	ret

instead of:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	movd	%xmm0, %rax
	movd	%eax, %xmm0
	shrq	$32, %rax
	movd	%eax, %xmm2
	addss	%xmm0, %xmm2
	movd	%xmm1, %rax
	movd	%eax, %xmm1
	addss	%xmm2, %xmm1
	shrq	$32, %rax
	movd	%eax, %xmm0
	addss	%xmm1, %xmm0
	ret

... eliminating half of the horribleness.

llvm-svn: 112227

2 files changed, 56 insertions, 0 deletions

diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
index dbe5200d000..ef5bbc4798d 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
@@ -1335,6 +1335,35 @@ static Instruction *OptimizeVectorResize(Value *InVal, const VectorType *DestTy,
   return new ShuffleVectorInst(InVal, V2, Mask);
 }
 
+/// OptimizeIntToFloatBitCast - See if we can optimize an integer->float/double
+/// bitcast.  The various long double bitcasts can't get in here.
+static Instruction *OptimizeIntToFloatBitCast(BitCastInst &CI,InstCombiner &IC) {
+  Value *Src = CI.getOperand(0);
+
+  // If this is a bitcast from int to float, check to see if the int is an
+  // extraction from a vector.
+  Value *VecInput = 0;
+  if (match(Src, m_Trunc(m_BitCast(m_Value(VecInput)))) &&
+      isa<VectorType>(VecInput->getType())) {
+    const VectorType *VecTy = cast<VectorType>(VecInput->getType());
+    const Type *DestTy = CI.getType();
+    
+    // If the element type of the vector doesn't match the result type, but the
+    // vector type's size is a multiple of the result type, bitcast it to be a
+    // vector type we can extract from.
+    if (VecTy->getElementType() != DestTy &&
+        VecTy->getPrimitiveSizeInBits() % DestTy->getPrimitiveSizeInBits()==0) {
+      VecTy = VectorType::get(DestTy,
+            VecTy->getPrimitiveSizeInBits() / DestTy->getPrimitiveSizeInBits());
+      VecInput = IC.Builder->CreateBitCast(VecInput, VecTy);
+    }
+    
+    if (VecTy->getElementType() == DestTy)
+      return ExtractElementInst::Create(VecInput, IC.Builder->getInt32(0));
+  }
+  
+  return 0;
+}
 
 Instruction *InstCombiner::visitBitCast(BitCastInst &CI) {
   // If the operands are integer typed then apply the integer transforms,
@@ -1386,6 +1415,11 @@ Instruction *InstCombiner::visitBitCast(BitCastInst &CI) {
                                                ((Instruction*)NULL));
     }
   }
+  
+  // Try to optimize int -> float bitcasts.
+  if ((DestTy->isFloatTy() || DestTy->isDoubleTy()) && isa<IntegerType>(SrcTy))
+    if (Instruction *I = OptimizeIntToFloatBitCast(CI, *this))
+      return I;
 
   if (const VectorType *DestVTy = dyn_cast<VectorType>(DestTy)) {
     if (DestVTy->getNumElements() == 1 && !SrcTy->isVectorTy()) {
diff --git a/llvm/test/Transforms/InstCombine/bitcast.ll b/llvm/test/Transforms/InstCombine/bitcast.ll
index 88fa9a43e54..c248b5e4bc8 100644
--- a/llvm/test/Transforms/InstCombine/bitcast.ll
+++ b/llvm/test/Transforms/InstCombine/bitcast.ll
@@ -13,3 +13,25 @@ define i32 @test1(i64 %a) {
 ; CHECK: ret i32 0
 }
 
+; Optimize bitcasts that are extracting low element of vector.  This happens
+; because of SRoA.
+; rdar://7892780
+define float @test2(<2 x float> %A, <2 x i32> %B) {
+  %tmp28 = bitcast <2 x float> %A to i64  ; <i64> [#uses=2]
+  %tmp23 = trunc i64 %tmp28 to i32                ; <i32> [#uses=1]
+  %tmp24 = bitcast i32 %tmp23 to float            ; <float> [#uses=1]
+
+  %tmp = bitcast <2 x i32> %B to i64
+  %tmp2 = trunc i64 %tmp to i32                ; <i32> [#uses=1]
+  %tmp4 = bitcast i32 %tmp2 to float            ; <float> [#uses=1]
+
+  %add = fadd float %tmp24, %tmp4
+  ret float %add
+  
+; CHECK: @test2
+; CHECK-NEXT:  %tmp24 = extractelement <2 x float> %A, i32 0
+; CHECK-NEXT:  bitcast <2 x i32> %B to <2 x float>
+; CHECK-NEXT:  %tmp4 = extractelement <2 x float> {{.*}}, i32 0
+; CHECK-NEXT:  %add = fadd float %tmp24, %tmp4
+; CHECK-NEXT:  ret float %add
+}