1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
llvm-mirror/test/CodeGen/AArch64/arm64-vector-insertion.ll
Evandro Menezes a4f6125cf5 [AArch64] Improve code generation of vector build
Instead of using, for example, `dup v0.4s, wzr`, which transfers between
register files, use the more efficient `movi v0.4s, #0` instead.

Differential revision: https://reviews.llvm.org/D41515

llvm-svn: 321824
2018-01-04 21:43:12 +00:00

33 lines
1.0 KiB
LLVM

; RUN: llc < %s -mtriple=arm64-eabi -mcpu=generic -aarch64-neon-syntax=apple | FileCheck %s
define void @test0f(float* nocapture %x, float %a) #0 {
entry:
%0 = insertelement <4 x float> <float undef, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00>, float %a, i32 0
%1 = bitcast float* %x to <4 x float>*
store <4 x float> %0, <4 x float>* %1, align 16
ret void
; CHECK-LABEL: test0f
; CHECK: movi.4s v[[TEMP:[0-9]+]], #0
; CHECK: mov.s v[[TEMP]][0], v{{[0-9]+}}[0]
; CHECK: str q[[TEMP]], [x0]
; CHECK: ret
}
define void @test1f(float* nocapture %x, float %a) #0 {
entry:
%0 = insertelement <4 x float> <float undef, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, float %a, i32 0
%1 = bitcast float* %x to <4 x float>*
store <4 x float> %0, <4 x float>* %1, align 16
ret void
; CHECK-LABEL: test1f
; CHECK: fmov s[[TEMP:[0-9]+]], #1.0000000
; CHECK: dup.4s v[[TEMP2:[0-9]+]], v[[TEMP]][0]
; CHECK: mov.s v[[TEMP2]][0], v0[0]
; CHECK: str q[[TEMP2]], [x0]
; CHECK: ret
}