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
|
/* Wrapper implementations of vector math functions.
Copyright (C) 2014-2023 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<https://www.gnu.org/licenses/>. */
/* SSE2 ISA version as wrapper to scalar. */
.macro WRAPPER_IMPL_SSE2 callee
push %rbx
cfi_adjust_cfa_offset (8)
cfi_rel_offset (%rbx, 0)
subq $16, %rsp
cfi_adjust_cfa_offset (16)
movaps %xmm0, (%rsp)
call JUMPTARGET(\callee)
movss %xmm0, (%rsp)
movss 4(%rsp), %xmm0
call JUMPTARGET(\callee)
movss %xmm0, 4(%rsp)
movss 8(%rsp), %xmm0
call JUMPTARGET(\callee)
movd %xmm0, %ebx
movss 12(%rsp), %xmm0
call JUMPTARGET(\callee)
movd %ebx, %xmm1
unpcklps %xmm0, %xmm1
movsd (%rsp), %xmm0
unpcklpd %xmm1, %xmm0
addq $16, %rsp
cfi_adjust_cfa_offset (-16)
popq %rbx
cfi_adjust_cfa_offset (-8)
cfi_restore (%rbx)
ret
.endm
/* 2 argument SSE2 ISA version as wrapper to scalar. */
.macro WRAPPER_IMPL_SSE2_ff callee
push %rbx
cfi_adjust_cfa_offset (8)
cfi_rel_offset (%rbx, 0)
subq $32, %rsp
cfi_adjust_cfa_offset (40)
movaps %xmm0, (%rsp)
movaps %xmm1, 16(%rsp)
call JUMPTARGET(\callee)
movss 20(%rsp), %xmm1
movss %xmm0, 0(%rsp)
movss 4(%rsp), %xmm0
call JUMPTARGET(\callee)
movss 24(%rsp), %xmm1
movss %xmm0, 4(%rsp)
movss 8(%rsp), %xmm0
call JUMPTARGET(\callee)
movss 28(%rsp), %xmm1
movd %xmm0, %ebx
movss 12(%rsp), %xmm0
call JUMPTARGET(\callee)
/* merge 4x results into xmm0. */
movd %ebx, %xmm1
unpcklps %xmm0, %xmm1
movsd (%rsp), %xmm0
unpcklpd %xmm1, %xmm0
addq $32, %rsp
cfi_adjust_cfa_offset (-32)
popq %rbx
cfi_adjust_cfa_offset (-8)
cfi_restore (%rbx)
ret
.endm
/* 3 argument SSE2 ISA version as wrapper to scalar. */
.macro WRAPPER_IMPL_SSE2_fFF callee
pushq %rbp
cfi_adjust_cfa_offset (8)
cfi_rel_offset (%rbp, 0)
pushq %rbx
cfi_adjust_cfa_offset (8)
cfi_rel_offset (%rbx, 0)
movq %rdi, %rbp
movq %rsi, %rbx
subq $24, %rsp
cfi_adjust_cfa_offset (24)
movaps %xmm0, (%rsp)
call JUMPTARGET(\callee)
movss 4(%rsp), %xmm0
leaq 4(%rbp), %rdi
leaq 4(%rbx), %rsi
call JUMPTARGET(\callee)
movss 8(%rsp), %xmm0
leaq 8(%rbp), %rdi
leaq 8(%rbx), %rsi
call JUMPTARGET(\callee)
movss 12(%rsp), %xmm0
leaq 12(%rbp), %rdi
leaq 12(%rbx), %rsi
call JUMPTARGET(\callee)
addq $24, %rsp
cfi_adjust_cfa_offset (-24)
popq %rbx
cfi_adjust_cfa_offset (-8)
cfi_restore (%rbx)
popq %rbp
cfi_adjust_cfa_offset (-8)
cfi_restore (%rbp)
ret
.endm
#include "svml_sd_wrapper_impl.h"
|