mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-12-08 12:08:46 +00:00
701b911eac
Use the whole gammut of constant immediates available to set up a vector. Instead of using, for example, `mov w0, #0xffff; dup v0.4s, w0`, which transfers between register files, use the more efficient `movi v0.4s, #-1` instead. Not limited to just a few values, but any immediate value that can be encoded by all the variants of `FMOV`, `MOVI`, `MVNI`, thus eliminating the need to there be patterns to optimize special cases. Differential revision: https://reviews.llvm.org/D42133 llvm-svn: 326718
32 lines
985 B
LLVM
32 lines
985 B
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.2d 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.4s v[[TEMP:[0-9]+]], #1.0
|
|
; CHECK: mov.s v[[TEMP]][0], v0[0]
|
|
; CHECK: str q[[TEMP]], [x0]
|
|
; CHECK: ret
|
|
}
|