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 < %s -mtriple powerpc64le-unknown-linux | FileCheck %s
; RUN: llc < %s -mtriple powerpc64le-unknown-linux -debug-only=machine-scheduler \
; RUN: 2>&1 | FileCheck %s --check-prefix=LOG
; REQUIRES: DEBUG
define double @in_nostrict(double %a, double %b, double %c, double %d) {
; CHECK-LABEL: in_nostrict:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: mffs 0
; CHECK-NEXT: xsdivdp 1, 1, 2
; CHECK-NEXT: xsadddp 1, 1, 3
; CHECK-NEXT: xsadddp 0, 1, 0
; CHECK-NEXT: mtfsf 255, 4
; CHECK-NEXT: xsdivdp 1, 3, 4
; CHECK-NEXT: xsadddp 1, 1, 2
; CHECK-NEXT: xsadddp 1, 0, 1
; CHECK-NEXT: blr
;
; LOG: *** MI Scheduling ***
; LOG-NEXT: in_nostrict:%bb.0 entry
; LOG: ExitSU: MTFSF 255, %{{[0-9]+}}:f8rc, 0, 0
; LOG: *** MI Scheduling ***
; LOG-NEXT: in_nostrict:%bb.0 entry
; LOG: ExitSU: %{{[0-9]+}}:f8rc = MFFS implicit $rm
;
; LOG: *** MI Scheduling ***
; LOG-NEXT: in_nostrict:%bb.0 entry
; LOG: ExitSU: MTFSF 255, renamable $f{{[0-9]+}}, 0, 0
entry:
%0 = tail call double @llvm.ppc.readflm()
%1 = fdiv double %a, %b
%2 = fadd double %1, %c
%3 = fadd double %2, %0
call double @llvm.ppc.setflm(double %d)
%5 = fdiv double %c, %d
%6 = fadd double %5, %b
%7 = fadd double %3, %6
ret double %7
}
define double @in_strict(double %a, double %b, double %c, double %d) #0 {
; CHECK-LABEL: in_strict:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: mffs 0
; CHECK-NEXT: xsdivdp 1, 1, 2
; CHECK-NEXT: xsadddp 1, 1, 3
; CHECK-NEXT: xsadddp 0, 1, 0
; CHECK-NEXT: mtfsf 255, 4
; CHECK-NEXT: xsdivdp 1, 3, 4
; CHECK-NEXT: xsadddp 1, 1, 2
; CHECK-NEXT: xsadddp 1, 0, 1
; CHECK-NEXT: blr
;
; LOG: ***** MI Scheduling *****
; LOG-NEXT: in_strict:%bb.0 entry
; LOG: ExitSU: MTFSF 255, %{{[0-9]+}}:f8rc, 0, 0
; LOG: ***** MI Scheduling *****
; LOG-NEXT: in_strict:%bb.0 entry
; LOG: ExitSU: %{{[0-9]+}}:f8rc = MFFS implicit $rm
;
; LOG: ***** MI Scheduling *****
; LOG-NEXT: in_strict:%bb.0 entry
; LOG: ExitSU: MTFSF 255, renamable $f{{[0-9]+}}, 0, 0
entry:
%0 = tail call double @llvm.ppc.readflm()
%1 = call double @llvm.experimental.constrained.fdiv.f64(double %a, double %b, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
%2 = call double @llvm.experimental.constrained.fadd.f64(double %1, double %c, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
%3 = call double @llvm.experimental.constrained.fadd.f64(double %2, double %0, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
call double @llvm.ppc.setflm(double %d)
%5 = call double @llvm.experimental.constrained.fdiv.f64(double %c, double %d, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
%6 = call double @llvm.experimental.constrained.fadd.f64(double %5, double %b, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
%7 = call double @llvm.experimental.constrained.fadd.f64(double %3, double %6, metadata !"round.dynamic", metadata !"fpexcept.strict") #0
ret double %7
}
define void @cse_nomerge(double* %f1, double* %f2, double %f3) #0 {
; CHECK-LABEL: cse_nomerge:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: mflr 0
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset lr, 16
; CHECK-NEXT: .cfi_offset r30, -24
; CHECK-NEXT: .cfi_offset f31, -8
; CHECK-NEXT: std 30, -24(1) # 8-byte Folded Spill
; CHECK-NEXT: stfd 31, -8(1) # 8-byte Folded Spill
; CHECK-NEXT: std 0, 16(1)
; CHECK-NEXT: stdu 1, -64(1)
; CHECK-NEXT: fmr 31, 1
; CHECK-NEXT: mr 30, 4
; CHECK-NEXT: mffs 0
; CHECK-NEXT: stfd 0, 0(3)
; CHECK-NEXT: bl effect_func
; CHECK-NEXT: nop
; CHECK-NEXT: mffs 0
; CHECK-NEXT: stfd 0, 0(30)
; CHECK-NEXT: mtfsf 255, 31
; CHECK-NEXT: addi 1, 1, 64
; CHECK-NEXT: ld 0, 16(1)
; CHECK-NEXT: lfd 31, -8(1) # 8-byte Folded Reload
; CHECK-NEXT: ld 30, -24(1) # 8-byte Folded Reload
; CHECK-NEXT: mtlr 0
; CHECK-NEXT: blr
entry:
%0 = call double @llvm.ppc.readflm()
store double %0, double* %f1, align 8
call void @effect_func()
%1 = call double @llvm.ppc.readflm()
store double %1, double* %f2, align 8
%2 = call contract double @llvm.ppc.setflm(double %f3)
ret void
}
define void @cse_nomerge_readonly(double* %f1, double* %f2, double %f3) #0 {
; CHECK-LABEL: cse_nomerge_readonly:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: mflr 0
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset lr, 16
; CHECK-NEXT: .cfi_offset r30, -24
; CHECK-NEXT: .cfi_offset f31, -8
; CHECK-NEXT: std 30, -24(1) # 8-byte Folded Spill
; CHECK-NEXT: stfd 31, -8(1) # 8-byte Folded Spill
; CHECK-NEXT: std 0, 16(1)
; CHECK-NEXT: stdu 1, -64(1)
; CHECK-NEXT: fmr 31, 1
; CHECK-NEXT: mr 30, 4
; CHECK-NEXT: mffs 0
; CHECK-NEXT: stfd 0, 0(3)
; CHECK-NEXT: bl readonly_func
; CHECK-NEXT: nop
; CHECK-NEXT: mffs 0
; CHECK-NEXT: stfd 0, 0(30)
; CHECK-NEXT: mtfsf 255, 31
; CHECK-NEXT: addi 1, 1, 64
; CHECK-NEXT: ld 0, 16(1)
; CHECK-NEXT: lfd 31, -8(1) # 8-byte Folded Reload
; CHECK-NEXT: ld 30, -24(1) # 8-byte Folded Reload
; CHECK-NEXT: mtlr 0
; CHECK-NEXT: blr
entry:
%0 = call double @llvm.ppc.readflm()
store double %0, double* %f1, align 8
call void @readonly_func()
%1 = call double @llvm.ppc.readflm()
store double %1, double* %f2, align 8
%2 = call contract double @llvm.ppc.setflm(double %f3)
ret void
}
declare void @effect_func()
declare void @readonly_func() #1
declare double @llvm.ppc.readflm()
declare double @llvm.ppc.setflm(double)
declare double @llvm.experimental.constrained.fadd.f64(double, double, metadata, metadata)
declare double @llvm.experimental.constrained.fdiv.f64(double, double, metadata, metadata)
attributes #0 = { strictfp }
attributes #1 = { readonly }
|
