834cc88c5d
lowerBuildVectorAsBroadcast will not broadcast splat constants in all cases, resulting in a lot of situations where a full width vector load that has failed to fold but is loading splat constant values could use a broadcast load instruction just as cheaply, and save constant pool space. NOTE: SSE3 targets can use MOVDDUP but not all SSE era CPUs can perform this as cheaply as a vector load, we will need to add scheduler model checks if we want to pursue this. This is an updated commit of 98061013e01207444cfd3980cde17b5e75764fbe after being reverted at a279a09ab9524d1d74ef29b34618102d4b202e2f
17 lines
638 B
LLVM
17 lines
638 B
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx | FileCheck %s
|
|
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx | FileCheck %s
|
|
|
|
; Check that constant integers are correctly being truncated before float conversion
|
|
|
|
define <4 x float> @test1() {
|
|
; CHECK-LABEL: test1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vmovddup {{.*#+}} xmm0 = [-1.0E+0,0.0E+0,-1.0E+0,0.0E+0]
|
|
; CHECK-NEXT: # xmm0 = mem[0,0]
|
|
; CHECK-NEXT: ret{{[l|q]}}
|
|
%1 = trunc <4 x i3> <i3 -1, i3 -22, i3 7, i3 8> to <4 x i1>
|
|
%2 = sitofp <4 x i1> %1 to <4 x float>
|
|
ret <4 x float> %2
|
|
}
|