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
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s
define void @foo(<3 x float> %in, <4 x i8>* nocapture %out) nounwind {
; CHECK-LABEL: foo:
; CHECK: # BB#0:
; CHECK-NEXT: cvttps2dq %xmm0, %xmm0
; CHECK-NEXT: movl $255, %eax
; CHECK-NEXT: pinsrd $3, %eax, %xmm0
; CHECK-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
; CHECK-NEXT: movd %xmm0, (%rdi)
; CHECK-NEXT: retq
%t0 = fptoui <3 x float> %in to <3 x i8>
%t1 = shufflevector <3 x i8> %t0, <3 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 undef>
%t2 = insertelement <4 x i8> %t1, i8 -1, i32 3
store <4 x i8> %t2, <4 x i8>* %out, align 4
ret void
}
; Verify that the DAGCombiner doesn't wrongly fold a build_vector into a
; blend with a zero vector if the build_vector contains negative zero.
define <4 x float> @test_negative_zero_1(<4 x float> %A) {
; CHECK-LABEL: test_negative_zero_1:
; CHECK: # BB#0: # %entry
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[2],zero
; CHECK-NEXT: retq
entry:
%0 = extractelement <4 x float> %A, i32 0
%1 = insertelement <4 x float> undef, float %0, i32 0
%2 = insertelement <4 x float> %1, float -0.0, i32 1
%3 = extractelement <4 x float> %A, i32 2
%4 = insertelement <4 x float> %2, float %3, i32 2
%5 = insertelement <4 x float> %4, float 0.0, i32 3
ret <4 x float> %5
}
define <2 x double> @test_negative_zero_2(<2 x double> %A) {
; CHECK-LABEL: test_negative_zero_2:
; CHECK: # BB#0: # %entry
; CHECK-NEXT: movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
; CHECK-NEXT: retq
entry:
%0 = extractelement <2 x double> %A, i32 0
%1 = insertelement <2 x double> undef, double %0, i32 0
%2 = insertelement <2 x double> %1, double -0.0, i32 1
ret <2 x double> %2
}
define <4 x float> @test_buildvector_v4f32_register(float %f0, float %f1, float %f2, float %f3) {
; CHECK-LABEL: test_buildvector_v4f32_register:
; CHECK: # BB#0:
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm3[0]
; CHECK-NEXT: retq
%ins0 = insertelement <4 x float> undef, float %f0, i32 0
%ins1 = insertelement <4 x float> %ins0, float %f1, i32 1
%ins2 = insertelement <4 x float> %ins1, float %f2, i32 2
%ins3 = insertelement <4 x float> %ins2, float %f3, i32 3
ret <4 x float> %ins3
}
define <4 x float> @test_buildvector_v4f32_load(float* %p0, float* %p1, float* %p2, float* %p3) {
; CHECK-LABEL: test_buildvector_v4f32_load:
; CHECK: # BB#0:
; CHECK-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[2,3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],mem[0]
; CHECK-NEXT: retq
%f0 = load float, float* %p0, align 4
%f1 = load float, float* %p1, align 4
%f2 = load float, float* %p2, align 4
%f3 = load float, float* %p3, align 4
%ins0 = insertelement <4 x float> undef, float %f0, i32 0
%ins1 = insertelement <4 x float> %ins0, float %f1, i32 1
%ins2 = insertelement <4 x float> %ins1, float %f2, i32 2
%ins3 = insertelement <4 x float> %ins2, float %f3, i32 3
ret <4 x float> %ins3
}
define <4 x float> @test_buildvector_v4f32_partial_load(float %f0, float %f1, float %f2, float* %p3) {
; CHECK-LABEL: test_buildvector_v4f32_partial_load:
; CHECK: # BB#0:
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
; CHECK-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],mem[0]
; CHECK-NEXT: retq
%f3 = load float, float* %p3, align 4
%ins0 = insertelement <4 x float> undef, float %f0, i32 0
%ins1 = insertelement <4 x float> %ins0, float %f1, i32 1
%ins2 = insertelement <4 x float> %ins1, float %f2, i32 2
%ins3 = insertelement <4 x float> %ins2, float %f3, i32 3
ret <4 x float> %ins3
}
|
