1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-linux-gnu -mcpu=haswell | FileCheck %s
define void @WhileWithLoopInvariantOperation.21() {
; CHECK-LABEL: WhileWithLoopInvariantOperation.21:
; CHECK: # %bb.0: # %while.1.body.preheader
; CHECK-NEXT: movq (%rax), %rax
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT: vmovaps %xmm0, 32(%rax)
; CHECK-NEXT: vmovaps {{.*#+}} xmm0 = [4294967295,4294967295,0,0]
; CHECK-NEXT: vmaskmovps %ymm0, %ymm0, (%rax)
while.1.body.preheader:
%0 = load ptr, ptr undef, align 8, !invariant.load !0, !dereferenceable !1, !align !2
%1 = getelementptr inbounds i8, ptr %0, i64 32
tail call void @llvm.memset.p0.i64(ptr nonnull align 16 dereferenceable(16) %1, i8 0, i64 16, i1 false)
%2 = tail call <8 x float> @llvm.masked.load.v8f32.p0(ptr undef, i32 4, <8 x i1> <i1 true, i1 true, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <8 x float> <float undef, float undef, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00>)
%3 = tail call <16 x float> @llvm.masked.load.v16f32.p0(ptr undef, i32 4, <16 x i1> <i1 true, i1 true, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <16 x float> <float undef, float undef, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00>)
tail call void @llvm.masked.store.v16f32.p0(<16 x float> undef, ptr nonnull undef, i32 4, <16 x i1> <i1 true, i1 true, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>)
unreachable
}
declare void @llvm.memset.p0.i64(ptr nocapture writeonly, i8, i64, i1 immarg)
declare <8 x float> @llvm.masked.load.v8f32.p0(ptr, i32 immarg, <8 x i1>, <8 x float>)
declare <16 x float> @llvm.masked.load.v16f32.p0(ptr, i32 immarg, <16 x i1>, <16 x float>)
declare void @llvm.masked.store.v16f32.p0(<16 x float>, ptr, i32 immarg, <16 x i1>)
!0 = !{}
!1 = !{i64 65}
!2 = !{i64 16}
|
