Files
archived-llvm/test/CodeGen/X86/vec_insert-2.ll
Craig Topper d28a1eae08 [X86] Teach the execution domain fixing tables to use movlhps inplace of unpcklpd for the packed single domain.
MOVLHPS has a smaller encoding than UNPCKLPD in the legacy encodings. With VEX and EVEX encodings it doesn't matter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@313509 91177308-0d34-0410-b5e6-96231b3b80d8
2017-09-18 04:40:58 +00:00

69 lines
2.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X64
define <4 x float> @t1(float %s, <4 x float> %tmp) nounwind {
; X32-LABEL: t1:
; X32: # BB#0:
; X32-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[2,0]
; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
; X32-NEXT: retl
;
; X64-LABEL: t1:
; X64: # BB#0:
; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[2,0]
; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,0]
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: retq
%tmp1 = insertelement <4 x float> %tmp, float %s, i32 3
ret <4 x float> %tmp1
}
define <4 x i32> @t2(i32 %s, <4 x i32> %tmp) nounwind {
; X32-LABEL: t2:
; X32: # BB#0:
; X32-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[2,0]
; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
; X32-NEXT: retl
;
; X64-LABEL: t2:
; X64: # BB#0:
; X64-NEXT: movd %edi, %xmm1
; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[2,0]
; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
; X64-NEXT: retq
%tmp1 = insertelement <4 x i32> %tmp, i32 %s, i32 3
ret <4 x i32> %tmp1
}
define <2 x double> @t3(double %s, <2 x double> %tmp) nounwind {
; X32-LABEL: t3:
; X32: # BB#0:
; X32-NEXT: movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
; X32-NEXT: retl
;
; X64-LABEL: t3:
; X64: # BB#0:
; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: retq
%tmp1 = insertelement <2 x double> %tmp, double %s, i32 1
ret <2 x double> %tmp1
}
define <8 x i16> @t4(i16 %s, <8 x i16> %tmp) nounwind {
; X32-LABEL: t4:
; X32: # BB#0:
; X32-NEXT: pinsrw $5, {{[0-9]+}}(%esp), %xmm0
; X32-NEXT: retl
;
; X64-LABEL: t4:
; X64: # BB#0:
; X64-NEXT: pinsrw $5, %edi, %xmm0
; X64-NEXT: retq
%tmp1 = insertelement <8 x i16> %tmp, i16 %s, i32 5
ret <8 x i16> %tmp1
}