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
|
dnl AMD64 mpn_addlshC_n -- rp[] = up[] + (vp[] << C)
dnl AMD64 mpn_rsblshC_n -- rp[] = (vp[] << C) - up[]
dnl Copyright 2009-2012 Free Software Foundation, Inc.
dnl This file is part of the GNU MP Library.
dnl
dnl The GNU MP Library is free software; you can redistribute it and/or modify
dnl it under the terms of either:
dnl
dnl * the GNU Lesser General Public License as published by the Free
dnl Software Foundation; either version 3 of the License, or (at your
dnl option) any later version.
dnl
dnl or
dnl
dnl * the GNU General Public License as published by the Free Software
dnl Foundation; either version 2 of the License, or (at your option) any
dnl later version.
dnl
dnl or both in parallel, as here.
dnl
dnl The GNU MP Library is distributed in the hope that it will be useful, but
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
dnl for more details.
dnl
dnl You should have received copies of the GNU General Public License and the
dnl GNU Lesser General Public License along with the GNU MP Library. If not,
dnl see https://www.gnu.org/licenses/.
C cycles/limb
C AMD K8,K9 ?
C AMD K10 ?
C Intel P4 ?
C Intel core2 3.25
C Intel NHM 4
C Intel SBR 2 C (or 1.95 when L(top)'s alignment = 16 (mod 32))
C Intel atom ?
C VIA nano ?
C This code probably runs close to optimally on Sandy Bridge (using 4-way
C unrolling). It also runs reasonably well on Core 2, but it runs poorly on
C all other processors, including Nehalem.
C INPUT PARAMETERS
define(`rp', `%rdi')
define(`up', `%rsi')
define(`vp', `%rdx')
define(`n', `%rcx')
define(`cy', `%r8')
ABI_SUPPORT(DOS64)
ABI_SUPPORT(STD64)
ASM_START()
TEXT
ALIGN(16)
PROLOGUE(func_nc)
FUNC_ENTRY(4)
IFDOS(` mov 56(%rsp), %r8 ')
push %rbp
mov cy, %rax
neg %rax C set msb on carry
xor R32(%rbp), R32(%rbp) C limb carry
mov (vp), %r8
shrd $RSH, %r8, %rbp
mov R32(n), R32(%r9)
and $3, R32(%r9)
je L(b00)
cmp $2, R32(%r9)
jc L(b01)
je L(b10)
jmp L(b11)
EPILOGUE()
ALIGN(16)
PROLOGUE(func_n)
FUNC_ENTRY(4)
push %rbp
xor R32(%rbp), R32(%rbp) C limb carry
mov (vp), %r8
shrd $RSH, %r8, %rbp
mov R32(n), R32(%rax)
and $3, R32(%rax)
je L(b00)
cmp $2, R32(%rax)
jc L(b01)
je L(b10)
L(b11): mov 8(vp), %r9
shrd $RSH, %r9, %r8
mov 16(vp), %r10
shrd $RSH, %r10, %r9
add R32(%rax), R32(%rax) C init carry flag
ADCSBB (up), %rbp
ADCSBB 8(up), %r8
ADCSBB 16(up), %r9
mov %rbp, (rp)
mov %r8, 8(rp)
mov %r9, 16(rp)
mov %r10, %rbp
lea 24(up), up
lea 24(vp), vp
lea 24(rp), rp
sbb R32(%rax), R32(%rax) C save carry flag
sub $3, n
ja L(top)
jmp L(end)
L(b01): add R32(%rax), R32(%rax) C init carry flag
ADCSBB (up), %rbp
mov %rbp, (rp)
mov %r8, %rbp
lea 8(up), up
lea 8(vp), vp
lea 8(rp), rp
sbb R32(%rax), R32(%rax) C save carry flag
sub $1, n
ja L(top)
jmp L(end)
L(b10): mov 8(vp), %r9
shrd $RSH, %r9, %r8
add R32(%rax), R32(%rax) C init carry flag
ADCSBB (up), %rbp
ADCSBB 8(up), %r8
mov %rbp, (rp)
mov %r8, 8(rp)
mov %r9, %rbp
lea 16(up), up
lea 16(vp), vp
lea 16(rp), rp
sbb R32(%rax), R32(%rax) C save carry flag
sub $2, n
ja L(top)
jmp L(end)
ALIGN(16)
L(top): mov (vp), %r8
shrd $RSH, %r8, %rbp
L(b00): mov 8(vp), %r9
shrd $RSH, %r9, %r8
mov 16(vp), %r10
shrd $RSH, %r10, %r9
mov 24(vp), %r11
shrd $RSH, %r11, %r10
lea 32(vp), vp
add R32(%rax), R32(%rax) C restore carry flag
ADCSBB (up), %rbp
ADCSBB 8(up), %r8
ADCSBB 16(up), %r9
ADCSBB 24(up), %r10
lea 32(up), up
mov %rbp, (rp)
mov %r8, 8(rp)
mov %r9, 16(rp)
mov %r10, 24(rp)
mov %r11, %rbp
lea 32(rp), rp
sbb R32(%rax), R32(%rax) C save carry flag
sub $4, n
jnz L(top)
L(end): shr $RSH, %rbp
add R32(%rax), R32(%rax) C restore carry flag
ADCSBB $0, %rbp
mov %rbp, %rax
pop %rbp
FUNC_EXIT()
ret
EPILOGUE()
|