• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s
3
4define <4 x i32> @foo(<4 x i32> %t, <4 x i32> %u) nounwind {
5; CHECK-LABEL: foo:
6; CHECK:       # %bb.0:
7; CHECK-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[3,3,3,3]
8; CHECK-NEXT:    movd %xmm2, %eax
9; CHECK-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[3,3,3,3]
10; CHECK-NEXT:    movd %xmm2, %ecx
11; CHECK-NEXT:    cltd
12; CHECK-NEXT:    idivl %ecx
13; CHECK-NEXT:    movd %edx, %xmm2
14; CHECK-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,3,2,3]
15; CHECK-NEXT:    movd %xmm3, %eax
16; CHECK-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,2,3]
17; CHECK-NEXT:    movd %xmm3, %ecx
18; CHECK-NEXT:    cltd
19; CHECK-NEXT:    idivl %ecx
20; CHECK-NEXT:    movd %edx, %xmm3
21; CHECK-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
22; CHECK-NEXT:    movd %xmm0, %eax
23; CHECK-NEXT:    movd %xmm1, %ecx
24; CHECK-NEXT:    cltd
25; CHECK-NEXT:    idivl %ecx
26; CHECK-NEXT:    movd %edx, %xmm2
27; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,1,1]
28; CHECK-NEXT:    movd %xmm0, %eax
29; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
30; CHECK-NEXT:    movd %xmm0, %ecx
31; CHECK-NEXT:    cltd
32; CHECK-NEXT:    idivl %ecx
33; CHECK-NEXT:    movd %edx, %xmm0
34; CHECK-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
35; CHECK-NEXT:    punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
36; CHECK-NEXT:    movdqa %xmm2, %xmm0
37; CHECK-NEXT:    retq
38  %m = srem <4 x i32> %t, %u
39  ret <4 x i32> %m
40}
41
42define <4 x i32> @bar(<4 x i32> %t, <4 x i32> %u) nounwind {
43; CHECK-LABEL: bar:
44; CHECK:       # %bb.0:
45; CHECK-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[3,3,3,3]
46; CHECK-NEXT:    movd %xmm2, %eax
47; CHECK-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[3,3,3,3]
48; CHECK-NEXT:    movd %xmm2, %ecx
49; CHECK-NEXT:    xorl %edx, %edx
50; CHECK-NEXT:    divl %ecx
51; CHECK-NEXT:    movd %edx, %xmm2
52; CHECK-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,3,2,3]
53; CHECK-NEXT:    movd %xmm3, %eax
54; CHECK-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,2,3]
55; CHECK-NEXT:    movd %xmm3, %ecx
56; CHECK-NEXT:    xorl %edx, %edx
57; CHECK-NEXT:    divl %ecx
58; CHECK-NEXT:    movd %edx, %xmm3
59; CHECK-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
60; CHECK-NEXT:    movd %xmm0, %eax
61; CHECK-NEXT:    movd %xmm1, %ecx
62; CHECK-NEXT:    xorl %edx, %edx
63; CHECK-NEXT:    divl %ecx
64; CHECK-NEXT:    movd %edx, %xmm2
65; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,1,1]
66; CHECK-NEXT:    movd %xmm0, %eax
67; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
68; CHECK-NEXT:    movd %xmm0, %ecx
69; CHECK-NEXT:    xorl %edx, %edx
70; CHECK-NEXT:    divl %ecx
71; CHECK-NEXT:    movd %edx, %xmm0
72; CHECK-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
73; CHECK-NEXT:    punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
74; CHECK-NEXT:    movdqa %xmm2, %xmm0
75; CHECK-NEXT:    retq
76  %m = urem <4 x i32> %t, %u
77  ret <4 x i32> %m
78}
79
80define <4 x float> @qux(<4 x float> %t, <4 x float> %u) nounwind {
81; CHECK-LABEL: qux:
82; CHECK:       # %bb.0:
83; CHECK-NEXT:    subq $72, %rsp
84; CHECK-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
85; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
86; CHECK-NEXT:    shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
87; CHECK-NEXT:    shufps {{.*#+}} xmm1 = xmm1[3,3,3,3]
88; CHECK-NEXT:    callq fmodf
89; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
90; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
91; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
92; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
93; CHECK-NEXT:    movhlps {{.*#+}} xmm1 = xmm1[1,1]
94; CHECK-NEXT:    callq fmodf
95; CHECK-NEXT:    unpcklps (%rsp), %xmm0 # 16-byte Folded Reload
96; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
97; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
98; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
99; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
100; CHECK-NEXT:    callq fmodf
101; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
102; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
103; CHECK-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
104; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
105; CHECK-NEXT:    shufps {{.*#+}} xmm1 = xmm1[1,1,1,1]
106; CHECK-NEXT:    callq fmodf
107; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
108; CHECK-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
109; CHECK-NEXT:    unpcklpd (%rsp), %xmm1 # 16-byte Folded Reload
110; CHECK-NEXT:    # xmm1 = xmm1[0],mem[0]
111; CHECK-NEXT:    movaps %xmm1, %xmm0
112; CHECK-NEXT:    addq $72, %rsp
113; CHECK-NEXT:    retq
114  %m = frem <4 x float> %t, %u
115  ret <4 x float> %m
116}
117