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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -mattr=-f16c | FileCheck %s -check-prefix=LIBCALL
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -mattr=+f16c | FileCheck %s -check-prefix=F16C
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -mattr=-f16c,+soft-float | FileCheck %s -check-prefix=SOFTFLOAT
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -mattr=+f16c,+soft-float | FileCheck %s -check-prefix=SOFTFLOAT
; This is a test for float to half float conversions on x86-64.
;
; If flag -soft-float is set, or if there is no F16C support, then:
; 1) half float to float conversions are
; translated into calls to __gnu_h2f_ieee defined
; by the compiler runtime library;
; 2) float to half float conversions are translated into calls
; to __gnu_f2h_ieee which expected to be defined by the
; compiler runtime library.
;
; Otherwise (we have F16C support):
; 1) half float to float conversion are translated using
; vcvtph2ps instructions;
; 2) float to half float conversions are translated using
; vcvtps2ph instructions
define void @test1(float %src, i16* %dest) {
; LIBCALL-LABEL: test1:
; LIBCALL: # %bb.0:
; LIBCALL-NEXT: pushq %rbx
; LIBCALL-NEXT: .cfi_def_cfa_offset 16
; LIBCALL-NEXT: .cfi_offset %rbx, -16
; LIBCALL-NEXT: movq %rdi, %rbx
; LIBCALL-NEXT: callq __gnu_f2h_ieee@PLT
; LIBCALL-NEXT: movw %ax, (%rbx)
; LIBCALL-NEXT: popq %rbx
; LIBCALL-NEXT: .cfi_def_cfa_offset 8
; LIBCALL-NEXT: retq
;
; F16C-LABEL: test1:
; F16C: # %bb.0:
; F16C-NEXT: vcvtps2ph $4, %xmm0, %xmm0
; F16C-NEXT: vpextrw $0, %xmm0, (%rdi)
; F16C-NEXT: retq
;
; SOFTFLOAT-LABEL: test1:
; SOFTFLOAT: # %bb.0:
; SOFTFLOAT-NEXT: pushq %rbx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 16
; SOFTFLOAT-NEXT: .cfi_offset %rbx, -16
; SOFTFLOAT-NEXT: movq %rsi, %rbx
; SOFTFLOAT-NEXT: callq __gnu_f2h_ieee@PLT
; SOFTFLOAT-NEXT: movw %ax, (%rbx)
; SOFTFLOAT-NEXT: popq %rbx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 8
; SOFTFLOAT-NEXT: retq
%1 = tail call i16 @llvm.convert.to.fp16.f32(float %src)
store i16 %1, i16* %dest, align 2
ret void
}
define float @test2(i16* nocapture %src) {
; LIBCALL-LABEL: test2:
; LIBCALL: # %bb.0:
; LIBCALL-NEXT: movzwl (%rdi), %edi
; LIBCALL-NEXT: jmp __gnu_h2f_ieee@PLT # TAILCALL
;
; F16C-LABEL: test2:
; F16C: # %bb.0:
; F16C-NEXT: movzwl (%rdi), %eax
; F16C-NEXT: vmovd %eax, %xmm0
; F16C-NEXT: vcvtph2ps %xmm0, %xmm0
; F16C-NEXT: retq
;
; SOFTFLOAT-LABEL: test2:
; SOFTFLOAT: # %bb.0:
; SOFTFLOAT-NEXT: pushq %rax
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 16
; SOFTFLOAT-NEXT: movzwl (%rdi), %edi
; SOFTFLOAT-NEXT: callq __gnu_h2f_ieee@PLT
; SOFTFLOAT-NEXT: popq %rcx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 8
; SOFTFLOAT-NEXT: retq
%1 = load i16, i16* %src, align 2
%2 = tail call float @llvm.convert.from.fp16.f32(i16 %1)
ret float %2
}
define float @test3(float %src) nounwind uwtable readnone {
; LIBCALL-LABEL: test3:
; LIBCALL: # %bb.0:
; LIBCALL-NEXT: pushq %rax
; LIBCALL-NEXT: .cfi_def_cfa_offset 16
; LIBCALL-NEXT: callq __gnu_f2h_ieee@PLT
; LIBCALL-NEXT: movzwl %ax, %edi
; LIBCALL-NEXT: popq %rax
; LIBCALL-NEXT: .cfi_def_cfa_offset 8
; LIBCALL-NEXT: jmp __gnu_h2f_ieee@PLT # TAILCALL
;
; F16C-LABEL: test3:
; F16C: # %bb.0:
; F16C-NEXT: vcvtps2ph $4, %xmm0, %xmm0
; F16C-NEXT: vcvtph2ps %xmm0, %xmm0
; F16C-NEXT: retq
;
; SOFTFLOAT-LABEL: test3:
; SOFTFLOAT: # %bb.0:
; SOFTFLOAT-NEXT: pushq %rax
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 16
; SOFTFLOAT-NEXT: callq __gnu_f2h_ieee@PLT
; SOFTFLOAT-NEXT: movzwl %ax, %edi
; SOFTFLOAT-NEXT: callq __gnu_h2f_ieee@PLT
; SOFTFLOAT-NEXT: popq %rcx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 8
; SOFTFLOAT-NEXT: retq
%1 = tail call i16 @llvm.convert.to.fp16.f32(float %src)
%2 = tail call float @llvm.convert.from.fp16.f32(i16 %1)
ret float %2
}
define double @test4(i16* nocapture %src) {
; LIBCALL-LABEL: test4:
; LIBCALL: # %bb.0:
; LIBCALL-NEXT: pushq %rax
; LIBCALL-NEXT: .cfi_def_cfa_offset 16
; LIBCALL-NEXT: movzwl (%rdi), %edi
; LIBCALL-NEXT: callq __gnu_h2f_ieee@PLT
; LIBCALL-NEXT: cvtss2sd %xmm0, %xmm0
; LIBCALL-NEXT: popq %rax
; LIBCALL-NEXT: .cfi_def_cfa_offset 8
; LIBCALL-NEXT: retq
;
; F16C-LABEL: test4:
; F16C: # %bb.0:
; F16C-NEXT: movzwl (%rdi), %eax
; F16C-NEXT: vmovd %eax, %xmm0
; F16C-NEXT: vcvtph2ps %xmm0, %xmm0
; F16C-NEXT: vcvtss2sd %xmm0, %xmm0, %xmm0
; F16C-NEXT: retq
;
; SOFTFLOAT-LABEL: test4:
; SOFTFLOAT: # %bb.0:
; SOFTFLOAT-NEXT: pushq %rax
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 16
; SOFTFLOAT-NEXT: movzwl (%rdi), %edi
; SOFTFLOAT-NEXT: callq __gnu_h2f_ieee@PLT
; SOFTFLOAT-NEXT: movl %eax, %edi
; SOFTFLOAT-NEXT: callq __extendsfdf2@PLT
; SOFTFLOAT-NEXT: popq %rcx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 8
; SOFTFLOAT-NEXT: retq
%1 = load i16, i16* %src, align 2
%2 = tail call double @llvm.convert.from.fp16.f64(i16 %1)
ret double %2
}
define i16 @test5(double %src) {
; LIBCALL-LABEL: test5:
; LIBCALL: # %bb.0:
; LIBCALL-NEXT: jmp __truncdfhf2@PLT # TAILCALL
;
; F16C-LABEL: test5:
; F16C: # %bb.0:
; F16C-NEXT: jmp __truncdfhf2@PLT # TAILCALL
;
; SOFTFLOAT-LABEL: test5:
; SOFTFLOAT: # %bb.0:
; SOFTFLOAT-NEXT: pushq %rax
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 16
; SOFTFLOAT-NEXT: callq __truncdfhf2@PLT
; SOFTFLOAT-NEXT: popq %rcx
; SOFTFLOAT-NEXT: .cfi_def_cfa_offset 8
; SOFTFLOAT-NEXT: retq
%val = tail call i16 @llvm.convert.to.fp16.f64(double %src)
ret i16 %val
}
declare float @llvm.convert.from.fp16.f32(i16) nounwind readnone
declare i16 @llvm.convert.to.fp16.f32(float) nounwind readnone
declare double @llvm.convert.from.fp16.f64(i16) nounwind readnone
declare i16 @llvm.convert.to.fp16.f64(double) nounwind readnone
|
