-; RUN: llc < %s -mcpu=generic -march=x86 -mattr=+sse4.2 | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse4.2 | FileCheck %s
define void @update(<3 x i8>* %dst, <3 x i8>* %src, i32 %n) nounwind {
+; CHECK-LABEL: update:
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: subl $12, %esp
+; CHECK-NEXT: movl $0, (%esp)
+; CHECK-NEXT: movdqa {{.*#+}} xmm0 = <1,1,1,u>
+; CHECK-NEXT: movdqa {{.*#+}} xmm1 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
+; CHECK-NEXT: jmp .LBB0_1
+; CHECK-NEXT: .p2align 4, 0x90
+; CHECK-NEXT: .LBB0_2: # %forbody
+; CHECK-NEXT: # in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT: movl (%esp), %eax
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %edx
+; CHECK-NEXT: pmovzxbd {{.*#+}} xmm2 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero
+; CHECK-NEXT: paddd %xmm0, %xmm2
+; CHECK-NEXT: pextrb $8, %xmm2, 2(%ecx,%eax,4)
+; CHECK-NEXT: pshufb %xmm1, %xmm2
+; CHECK-NEXT: pextrw $0, %xmm2, (%ecx,%eax,4)
+; CHECK-NEXT: incl (%esp)
+; CHECK-NEXT: .LBB0_1: # %forcond
+; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: movl (%esp), %eax
+; CHECK-NEXT: cmpl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: jl .LBB0_2
+; CHECK-NEXT: # BB#3: # %afterfor
+; CHECK-NEXT: addl $12, %esp
+; CHECK-NEXT: retl
entry:
-; CHECK-NOT: pextrw
-; CHECK: add
-
- %dst.addr = alloca <3 x i8>* ; <<3 x i8>**> [#uses=2]
- %src.addr = alloca <3 x i8>* ; <<3 x i8>**> [#uses=2]
- %n.addr = alloca i32 ; <i32*> [#uses=2]
- %i = alloca i32, align 4 ; <i32*> [#uses=6]
+ %dst.addr = alloca <3 x i8>*
+ %src.addr = alloca <3 x i8>*
+ %n.addr = alloca i32
+ %i = alloca i32, align 4
store <3 x i8>* %dst, <3 x i8>** %dst.addr
store <3 x i8>* %src, <3 x i8>** %src.addr
store i32 %n, i32* %n.addr
store i32 0, i32* %i
br label %forcond
-forcond: ; preds = %forinc, %entry
- %tmp = load i32, i32* %i ; <i32> [#uses=1]
- %tmp1 = load i32, i32* %n.addr ; <i32> [#uses=1]
- %cmp = icmp slt i32 %tmp, %tmp1 ; <i1> [#uses=1]
+forcond:
+ %tmp = load i32, i32* %i
+ %tmp1 = load i32, i32* %n.addr
+ %cmp = icmp slt i32 %tmp, %tmp1
br i1 %cmp, label %forbody, label %afterfor
-forbody: ; preds = %forcond
- %tmp2 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp3 = load <3 x i8>*, <3 x i8>** %dst.addr ; <<3 x i8>*> [#uses=1]
- %arrayidx = getelementptr <3 x i8>, <3 x i8>* %tmp3, i32 %tmp2 ; <<3 x i8>*> [#uses=1]
- %tmp4 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp5 = load <3 x i8>*, <3 x i8>** %src.addr ; <<3 x i8>*> [#uses=1]
- %arrayidx6 = getelementptr <3 x i8>, <3 x i8>* %tmp5, i32 %tmp4 ; <<3 x i8>*> [#uses=1]
- %tmp7 = load <3 x i8>, <3 x i8>* %arrayidx6 ; <<3 x i8>> [#uses=1]
- %add = add <3 x i8> %tmp7, < i8 1, i8 1, i8 1 > ; <<3 x i8>> [#uses=1]
+forbody:
+ %tmp2 = load i32, i32* %i
+ %tmp3 = load <3 x i8>*, <3 x i8>** %dst.addr
+ %arrayidx = getelementptr <3 x i8>, <3 x i8>* %tmp3, i32 %tmp2
+ %tmp4 = load i32, i32* %i
+ %tmp5 = load <3 x i8>*, <3 x i8>** %src.addr
+ %arrayidx6 = getelementptr <3 x i8>, <3 x i8>* %tmp5, i32 %tmp4
+ %tmp7 = load <3 x i8>, <3 x i8>* %arrayidx6
+ %add = add <3 x i8> %tmp7, < i8 1, i8 1, i8 1 >
store <3 x i8> %add, <3 x i8>* %arrayidx
br label %forinc
-forinc: ; preds = %forbody
- %tmp8 = load i32, i32* %i ; <i32> [#uses=1]
- %inc = add i32 %tmp8, 1 ; <i32> [#uses=1]
+forinc:
+ %tmp8 = load i32, i32* %i
+ %inc = add i32 %tmp8, 1
store i32 %inc, i32* %i
br label %forcond
-afterfor: ; preds = %forcond
+afterfor:
ret void
}
+
-; RUN: llc < %s -march=x86 -mattr=+sse4.2 | FileCheck %s
-; CHECK: padd
-; CHECK: pand
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse4.2 | FileCheck %s
; widen v8i8 to v16i8 (checks even power of 2 widening with add & and)
define void @update(i64* %dst_i, i64* %src_i, i32 %n) nounwind {
+; CHECK-LABEL: update:
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: subl $12, %esp
+; CHECK-NEXT: movl $0, (%esp)
+; CHECK-NEXT: movdqa {{.*#+}} xmm0 = [1,1,1,1,1,1,1,1]
+; CHECK-NEXT: movdqa {{.*#+}} xmm1 = [4,4,4,4,4,4,4,4]
+; CHECK-NEXT: movdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
+; CHECK-NEXT: jmp .LBB0_1
+; CHECK-NEXT: .p2align 4, 0x90
+; CHECK-NEXT: .LBB0_2: # %forbody
+; CHECK-NEXT: # in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT: movl (%esp), %eax
+; CHECK-NEXT: shll $3, %eax
+; CHECK-NEXT: addl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; CHECK-NEXT: movl (%esp), %eax
+; CHECK-NEXT: shll $3, %eax
+; CHECK-NEXT: addl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; CHECK-NEXT: movl (%esp), %ecx
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %edx
+; CHECK-NEXT: pmovzxbw {{.*#+}} xmm3 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
+; CHECK-NEXT: paddw %xmm0, %xmm3
+; CHECK-NEXT: pand %xmm1, %xmm3
+; CHECK-NEXT: pshufb %xmm2, %xmm3
+; CHECK-NEXT: movq %xmm3, (%edx,%ecx,8)
+; CHECK-NEXT: incl (%esp)
+; CHECK-NEXT: .LBB0_1: # %forcond
+; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: movl (%esp), %eax
+; CHECK-NEXT: cmpl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: jl .LBB0_2
+; CHECK-NEXT: # BB#3: # %afterfor
+; CHECK-NEXT: addl $12, %esp
+; CHECK-NEXT: retl
entry:
- %dst_i.addr = alloca i64* ; <i64**> [#uses=2]
- %src_i.addr = alloca i64* ; <i64**> [#uses=2]
- %n.addr = alloca i32 ; <i32*> [#uses=2]
- %i = alloca i32, align 4 ; <i32*> [#uses=8]
- %dst = alloca <8 x i8>*, align 4 ; <<8 x i8>**> [#uses=2]
- %src = alloca <8 x i8>*, align 4 ; <<8 x i8>**> [#uses=2]
+ %dst_i.addr = alloca i64*
+ %src_i.addr = alloca i64*
+ %n.addr = alloca i32
+ %i = alloca i32, align 4
+ %dst = alloca <8 x i8>*, align 4
+ %src = alloca <8 x i8>*, align 4
store i64* %dst_i, i64** %dst_i.addr
store i64* %src_i, i64** %src_i.addr
store i32 %n, i32* %n.addr
store i32 0, i32* %i
br label %forcond
-forcond: ; preds = %forinc, %entry
- %tmp = load i32, i32* %i ; <i32> [#uses=1]
- %tmp1 = load i32, i32* %n.addr ; <i32> [#uses=1]
- %cmp = icmp slt i32 %tmp, %tmp1 ; <i1> [#uses=1]
+forcond:
+ %tmp = load i32, i32* %i
+ %tmp1 = load i32, i32* %n.addr
+ %cmp = icmp slt i32 %tmp, %tmp1
br i1 %cmp, label %forbody, label %afterfor
-forbody: ; preds = %forcond
- %tmp2 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp3 = load i64*, i64** %dst_i.addr ; <i64*> [#uses=1]
- %arrayidx = getelementptr i64, i64* %tmp3, i32 %tmp2 ; <i64*> [#uses=1]
- %conv = bitcast i64* %arrayidx to <8 x i8>* ; <<8 x i8>*> [#uses=1]
+forbody:
+ %tmp2 = load i32, i32* %i
+ %tmp3 = load i64*, i64** %dst_i.addr
+ %arrayidx = getelementptr i64, i64* %tmp3, i32 %tmp2
+ %conv = bitcast i64* %arrayidx to <8 x i8>*
store <8 x i8>* %conv, <8 x i8>** %dst
- %tmp4 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp5 = load i64*, i64** %src_i.addr ; <i64*> [#uses=1]
- %arrayidx6 = getelementptr i64, i64* %tmp5, i32 %tmp4 ; <i64*> [#uses=1]
- %conv7 = bitcast i64* %arrayidx6 to <8 x i8>* ; <<8 x i8>*> [#uses=1]
+ %tmp4 = load i32, i32* %i
+ %tmp5 = load i64*, i64** %src_i.addr
+ %arrayidx6 = getelementptr i64, i64* %tmp5, i32 %tmp4
+ %conv7 = bitcast i64* %arrayidx6 to <8 x i8>*
store <8 x i8>* %conv7, <8 x i8>** %src
- %tmp8 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp9 = load <8 x i8>*, <8 x i8>** %dst ; <<8 x i8>*> [#uses=1]
- %arrayidx10 = getelementptr <8 x i8>, <8 x i8>* %tmp9, i32 %tmp8 ; <<8 x i8>*> [#uses=1]
- %tmp11 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp12 = load <8 x i8>*, <8 x i8>** %src ; <<8 x i8>*> [#uses=1]
- %arrayidx13 = getelementptr <8 x i8>, <8 x i8>* %tmp12, i32 %tmp11 ; <<8 x i8>*> [#uses=1]
- %tmp14 = load <8 x i8>, <8 x i8>* %arrayidx13 ; <<8 x i8>> [#uses=1]
- %add = add <8 x i8> %tmp14, < i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1 > ; <<8 x i8>> [#uses=1]
- %and = and <8 x i8> %add, < i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4 > ; <<8 x i8>> [#uses=1]
+ %tmp8 = load i32, i32* %i
+ %tmp9 = load <8 x i8>*, <8 x i8>** %dst
+ %arrayidx10 = getelementptr <8 x i8>, <8 x i8>* %tmp9, i32 %tmp8
+ %tmp11 = load i32, i32* %i
+ %tmp12 = load <8 x i8>*, <8 x i8>** %src
+ %arrayidx13 = getelementptr <8 x i8>, <8 x i8>* %tmp12, i32 %tmp11
+ %tmp14 = load <8 x i8>, <8 x i8>* %arrayidx13
+ %add = add <8 x i8> %tmp14, < i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1 >
+ %and = and <8 x i8> %add, < i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4 >
store <8 x i8> %and, <8 x i8>* %arrayidx10
br label %forinc
-forinc: ; preds = %forbody
- %tmp15 = load i32, i32* %i ; <i32> [#uses=1]
- %inc = add i32 %tmp15, 1 ; <i32> [#uses=1]
+forinc:
+ %tmp15 = load i32, i32* %i
+ %inc = add i32 %tmp15, 1
store i32 %inc, i32* %i
br label %forcond
-afterfor: ; preds = %forcond
+afterfor:
ret void
}
-; RUN: llc < %s -mcpu=generic -march=x86 -mattr=+sse4.2 -post-RA-scheduler=true | FileCheck %s
-; CHECK: paddd
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse4.2 -post-RA-scheduler=true | FileCheck %s
; Widen a v3i16 to v8i16 to do a vector add
-@.str = internal constant [4 x i8] c"%d \00" ; <[4 x i8]*> [#uses=1]
-@.str1 = internal constant [2 x i8] c"\0A\00" ; <[2 x i8]*> [#uses=1]
+@.str = internal constant [4 x i8] c"%d \00"
+@.str1 = internal constant [2 x i8] c"\0A\00"
define void @update(<3 x i16>* %dst, <3 x i16>* %src, i32 %n) nounwind {
+; CHECK-LABEL: update:
+; CHECK: # BB#0: # %entry
+; CHECK-NEXT: pushl %ebp
+; CHECK-NEXT: movl %esp, %ebp
+; CHECK-NEXT: andl $-8, %esp
+; CHECK-NEXT: subl $40, %esp
+; CHECK-NEXT: movl {{\.LCPI.*}}, %eax
+; CHECK-NEXT: movdqa {{.*#+}} xmm0 = <1,1,1,u>
+; CHECK-NEXT: movdqa {{.*#+}} xmm1 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; CHECK-NEXT: movl $0, {{[0-9]+}}(%esp)
+; CHECK-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; CHECK-NEXT: movw $1, {{[0-9]+}}(%esp)
+; CHECK-NEXT: jmp .LBB0_1
+; CHECK-NEXT: .p2align 4, 0x90
+; CHECK-NEXT: .LBB0_2: # %forbody
+; CHECK-NEXT: # in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: movl 12(%ebp), %edx
+; CHECK-NEXT: movl 8(%ebp), %ecx
+; CHECK-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; CHECK-NEXT: pmovzxwd {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero
+; CHECK-NEXT: pinsrd $2, 4(%edx,%eax,8), %xmm2
+; CHECK-NEXT: paddd %xmm0, %xmm2
+; CHECK-NEXT: pextrw $4, %xmm2, 4(%ecx,%eax,8)
+; CHECK-NEXT: pshufb %xmm1, %xmm2
+; CHECK-NEXT: movd %xmm2, (%ecx,%eax,8)
+; CHECK-NEXT: incl {{[0-9]+}}(%esp)
+; CHECK-NEXT: .LBB0_1: # %forcond
+; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
+; CHECK-NEXT: cmpl 16(%ebp), %eax
+; CHECK-NEXT: jl .LBB0_2
+; CHECK-NEXT: # BB#3: # %afterfor
+; CHECK-NEXT: movl %ebp, %esp
+; CHECK-NEXT: popl %ebp
+; CHECK-NEXT: retl
entry:
- %dst.addr = alloca <3 x i16>* ; <<3 x i16>**> [#uses=2]
- %src.addr = alloca <3 x i16>* ; <<3 x i16>**> [#uses=2]
- %n.addr = alloca i32 ; <i32*> [#uses=2]
- %v = alloca <3 x i16>, align 8 ; <<3 x i16>*> [#uses=1]
- %i = alloca i32, align 4 ; <i32*> [#uses=6]
+ %dst.addr = alloca <3 x i16>*
+ %src.addr = alloca <3 x i16>*
+ %n.addr = alloca i32
+ %v = alloca <3 x i16>, align 8
+ %i = alloca i32, align 4
store <3 x i16>* %dst, <3 x i16>** %dst.addr
store <3 x i16>* %src, <3 x i16>** %src.addr
store i32 %n, i32* %n.addr
store i32 0, i32* %i
br label %forcond
-forcond: ; preds = %forinc, %entry
- %tmp = load i32, i32* %i ; <i32> [#uses=1]
- %tmp1 = load i32, i32* %n.addr ; <i32> [#uses=1]
- %cmp = icmp slt i32 %tmp, %tmp1 ; <i1> [#uses=1]
+forcond:
+ %tmp = load i32, i32* %i
+ %tmp1 = load i32, i32* %n.addr
+ %cmp = icmp slt i32 %tmp, %tmp1
br i1 %cmp, label %forbody, label %afterfor
-forbody: ; preds = %forcond
- %tmp2 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp3 = load <3 x i16>*, <3 x i16>** %dst.addr ; <<3 x i16>*> [#uses=1]
- %arrayidx = getelementptr <3 x i16>, <3 x i16>* %tmp3, i32 %tmp2 ; <<3 x i16>*> [#uses=1]
- %tmp4 = load i32, i32* %i ; <i32> [#uses=1]
- %tmp5 = load <3 x i16>*, <3 x i16>** %src.addr ; <<3 x i16>*> [#uses=1]
- %arrayidx6 = getelementptr <3 x i16>, <3 x i16>* %tmp5, i32 %tmp4 ; <<3 x i16>*> [#uses=1]
- %tmp7 = load <3 x i16>, <3 x i16>* %arrayidx6 ; <<3 x i16>> [#uses=1]
- %add = add <3 x i16> %tmp7, < i16 1, i16 1, i16 1 > ; <<3 x i16>> [#uses=1]
+forbody:
+ %tmp2 = load i32, i32* %i
+ %tmp3 = load <3 x i16>*, <3 x i16>** %dst.addr
+ %arrayidx = getelementptr <3 x i16>, <3 x i16>* %tmp3, i32 %tmp2
+ %tmp4 = load i32, i32* %i
+ %tmp5 = load <3 x i16>*, <3 x i16>** %src.addr
+ %arrayidx6 = getelementptr <3 x i16>, <3 x i16>* %tmp5, i32 %tmp4
+ %tmp7 = load <3 x i16>, <3 x i16>* %arrayidx6
+ %add = add <3 x i16> %tmp7, < i16 1, i16 1, i16 1 >
store <3 x i16> %add, <3 x i16>* %arrayidx
br label %forinc
-forinc: ; preds = %forbody
- %tmp8 = load i32, i32* %i ; <i32> [#uses=1]
- %inc = add i32 %tmp8, 1 ; <i32> [#uses=1]
+forinc:
+ %tmp8 = load i32, i32* %i
+ %inc = add i32 %tmp8, 1
store i32 %inc, i32* %i
br label %forcond
-afterfor: ; preds = %forcond
+afterfor:
ret void
}
projects / llvm / commitdiff