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
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -relocation-model=pic -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr9 < %s | FileCheck -check-prefix=CHECK-P9 %s
; RUN: llc -relocation-model=pic -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr8 < %s | FileCheck -check-prefix=CHECK-P8 %s
; As the constant could be represented as float, a float is
; loaded from constant pool.
define double @doubleConstant1() {
; CHECK-P9-LABEL: doubleConstant1:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: vspltisw 2, 14
; CHECK-P9-NEXT: xvcvsxwdp 1, 34
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: doubleConstant1:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: vspltisw 2, 14
; CHECK-P8-NEXT: xvcvsxwdp 1, 34
; CHECK-P8-NEXT: blr
ret double 1.400000e+01
}
; As the constant couldn't be represented as float, a double is
; loaded from constant pool.
define double @doubleConstant2() {
; CHECK-P9-LABEL: doubleConstant2:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: addis 3, 2, .LCPI1_0@toc@ha
; CHECK-P9-NEXT: lfd 1, .LCPI1_0@toc@l(3)
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: doubleConstant2:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: addis 3, 2, .LCPI1_0@toc@ha
; CHECK-P8-NEXT: lfd 1, .LCPI1_0@toc@l(3)
; CHECK-P8-NEXT: blr
ret double 2.408904e+01
}
@FArr = hidden local_unnamed_addr global [10 x float] zeroinitializer, align 4
define float @floatConstantArray() local_unnamed_addr {
; CHECK-P9-LABEL: floatConstantArray:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: addis 3, 2, FArr@toc@ha+12
; CHECK-P9-NEXT: lfs 0, FArr@toc@l+12(3)
; CHECK-P9-NEXT: addis 3, 2, .LCPI2_0@toc@ha
; CHECK-P9-NEXT: lfs 1, .LCPI2_0@toc@l(3)
; CHECK-P9-NEXT: xsaddsp 1, 0, 1
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: floatConstantArray:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: addis 3, 2, FArr@toc@ha+12
; CHECK-P8-NEXT: lfs 0, FArr@toc@l+12(3)
; CHECK-P8-NEXT: addis 3, 2, .LCPI2_0@toc@ha
; CHECK-P8-NEXT: lfs 1, .LCPI2_0@toc@l(3)
; CHECK-P8-NEXT: xsaddsp 1, 0, 1
; CHECK-P8-NEXT: blr
%1 = load float, ptr getelementptr inbounds ([10 x float], ptr @FArr, i64 0, i64 3), align 4
%2 = fadd float %1, 0x400B333340000000
ret float %2
}
define float @floatConstant() {
; CHECK-P9-LABEL: floatConstant:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: addis 3, 2, .LCPI3_0@toc@ha
; CHECK-P9-NEXT: lfs 1, .LCPI3_0@toc@l(3)
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: floatConstant:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: addis 3, 2, .LCPI3_0@toc@ha
; CHECK-P8-NEXT: lfs 1, .LCPI3_0@toc@l(3)
; CHECK-P8-NEXT: blr
ret float 0x400470A3E0000000
}
; llvm put the hidden globals into the TOC table.
; TODO - do some analysis and decide which globals could be put into TOC.
@d = hidden local_unnamed_addr global [200 x double] zeroinitializer, align 8
define double @doubleConstantArray() {
; CHECK-P9-LABEL: doubleConstantArray:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: addis 3, 2, d@toc@ha+24
; CHECK-P9-NEXT: lfd 0, d@toc@l+24(3)
; CHECK-P9-NEXT: addis 3, 2, .LCPI4_0@toc@ha
; CHECK-P9-NEXT: lfd 1, .LCPI4_0@toc@l(3)
; CHECK-P9-NEXT: xsadddp 1, 0, 1
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: doubleConstantArray:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: addis 3, 2, d@toc@ha+24
; CHECK-P8-NEXT: lfd 0, d@toc@l+24(3)
; CHECK-P8-NEXT: addis 3, 2, .LCPI4_0@toc@ha
; CHECK-P8-NEXT: lfd 1, .LCPI4_0@toc@l(3)
; CHECK-P8-NEXT: xsadddp 1, 0, 1
; CHECK-P8-NEXT: blr
%1 = load double, ptr getelementptr inbounds ([200 x double], ptr @d, i64 0, i64 3), align 8
%2 = fadd double %1, 6.880000e+00
ret double %2
}
@arr = hidden local_unnamed_addr global [20000 x double] zeroinitializer, align 8
; Access an element with an offset that doesn't fit in the displacement field of LFD.
define double @doubleLargeConstantArray() {
; CHECK-P9-LABEL: doubleLargeConstantArray:
; CHECK-P9: # %bb.0:
; CHECK-P9-NEXT: addis 3, 2, arr@toc@ha
; CHECK-P9-NEXT: li 4, 0
; CHECK-P9-NEXT: addi 3, 3, arr@toc@l
; CHECK-P9-NEXT: ori 4, 4, 32768
; CHECK-P9-NEXT: lfdx 0, 3, 4
; CHECK-P9-NEXT: addis 3, 2, .LCPI5_0@toc@ha
; CHECK-P9-NEXT: lfd 1, .LCPI5_0@toc@l(3)
; CHECK-P9-NEXT: xsadddp 1, 0, 1
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: doubleLargeConstantArray:
; CHECK-P8: # %bb.0:
; CHECK-P8-NEXT: addis 3, 2, arr@toc@ha
; CHECK-P8-NEXT: li 4, 0
; CHECK-P8-NEXT: addi 3, 3, arr@toc@l
; CHECK-P8-NEXT: ori 4, 4, 32768
; CHECK-P8-NEXT: lfdx 0, 3, 4
; CHECK-P8-NEXT: addis 3, 2, .LCPI5_0@toc@ha
; CHECK-P8-NEXT: lfd 1, .LCPI5_0@toc@l(3)
; CHECK-P8-NEXT: xsadddp 1, 0, 1
; CHECK-P8-NEXT: blr
%1 = load double, ptr getelementptr inbounds ([20000 x double], ptr @arr, i64 0, i64 4096), align 8
%2 = fadd double %1, 6.880000e+00
ret double %2
}
@vec_arr = global [10 x <4 x i32>] zeroinitializer, align 16
define <4 x i32> @vectorArray() #0 {
; CHECK-P9-LABEL: vectorArray:
; CHECK-P9: # %bb.0: # %entry
; CHECK-P9-NEXT: addis 3, 2, .LC0@toc@ha
; CHECK-P9-NEXT: ld 3, .LC0@toc@l(3)
; CHECK-P9-NEXT: lxv 34, 32(3)
; CHECK-P9-NEXT: blr
;
; CHECK-P8-LABEL: vectorArray:
; CHECK-P8: # %bb.0: # %entry
; CHECK-P8-NEXT: addis 3, 2, .LC0@toc@ha
; CHECK-P8-NEXT: ld 3, .LC0@toc@l(3)
; CHECK-P8-NEXT: addi 3, 3, 32
; CHECK-P8-NEXT: lxvd2x 0, 0, 3
; CHECK-P8-NEXT: xxswapd 34, 0
; CHECK-P8-NEXT: blr
entry:
%0 = load <4 x i32>, ptr getelementptr inbounds ([10 x <4 x i32>], ptr @vec_arr, i64 0, i64 2), align 16
ret <4 x i32> %0
}
|
