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
|
dnl x86-64 mpn_rshift optimized for Pentium 4.
dnl Copyright 2003, 2005, 2007, 2008, 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/.
include(`../config.m4')
C cycles/limb
C AMD K8,K9 2.5
C AMD K10 ?
C Intel P4 3.29
C Intel core2 2.1 (fluctuates, presumably cache related)
C Intel corei ?
C Intel atom 14.3
C VIA nano ?
C INPUT PARAMETERS
define(`rp',`%rdi')
define(`up',`%rsi')
define(`n',`%rdx')
define(`cnt',`%cl')
ABI_SUPPORT(DOS64)
ABI_SUPPORT(STD64)
ASM_START()
TEXT
ALIGN(32)
PROLOGUE(mpn_rshift)
FUNC_ENTRY(4)
mov (up), %rax
movd R32(%rcx), %mm4
neg R32(%rcx) C put lsh count in cl
and $63, R32(%rcx)
movd R32(%rcx), %mm5
lea -8(up,n,8), up
lea -8(rp,n,8), rp
lea 1(n), R32(%r8)
neg n
shl R8(%rcx), %rax C function return value
and $3, R32(%r8)
je L(rol) C jump for n = 3, 7, 11, ...
dec R32(%r8)
jne L(1)
C n = 4, 8, 12, ...
movq 8(up,n,8), %mm2
psrlq %mm4, %mm2
movq 16(up,n,8), %mm0
psllq %mm5, %mm0
por %mm0, %mm2
movq %mm2, 8(rp,n,8)
inc n
jmp L(rol)
L(1): dec R32(%r8)
je L(1x) C jump for n = 1, 5, 9, 13, ...
C n = 2, 6, 10, 16, ...
movq 8(up,n,8), %mm2
psrlq %mm4, %mm2
movq 16(up,n,8), %mm0
psllq %mm5, %mm0
por %mm0, %mm2
movq %mm2, 8(rp,n,8)
inc n
L(1x):
cmp $-1, n
je L(ast)
movq 8(up,n,8), %mm2
psrlq %mm4, %mm2
movq 16(up,n,8), %mm3
psrlq %mm4, %mm3
movq 16(up,n,8), %mm0
movq 24(up,n,8), %mm1
psllq %mm5, %mm0
por %mm0, %mm2
psllq %mm5, %mm1
por %mm1, %mm3
movq %mm2, 8(rp,n,8)
movq %mm3, 16(rp,n,8)
add $2, n
L(rol): movq 8(up,n,8), %mm2
psrlq %mm4, %mm2
movq 16(up,n,8), %mm3
psrlq %mm4, %mm3
add $4, n C 4
jb L(end) C 2
ALIGN(32)
L(top):
C finish stuff from lsh block
movq -16(up,n,8), %mm0
movq -8(up,n,8), %mm1
psllq %mm5, %mm0
por %mm0, %mm2
psllq %mm5, %mm1
movq (up,n,8), %mm0
por %mm1, %mm3
movq 8(up,n,8), %mm1
movq %mm2, -24(rp,n,8)
movq %mm3, -16(rp,n,8)
C start two new rsh
psllq %mm5, %mm0
psllq %mm5, %mm1
C finish stuff from rsh block
movq -8(up,n,8), %mm2
movq (up,n,8), %mm3
psrlq %mm4, %mm2
por %mm2, %mm0
psrlq %mm4, %mm3
movq 8(up,n,8), %mm2
por %mm3, %mm1
movq 16(up,n,8), %mm3
movq %mm0, -8(rp,n,8)
movq %mm1, (rp,n,8)
C start two new lsh
add $4, n
psrlq %mm4, %mm2
psrlq %mm4, %mm3
jae L(top) C 2
L(end):
movq -8(up), %mm0
psllq %mm5, %mm0
por %mm0, %mm2
movq (up), %mm1
psllq %mm5, %mm1
por %mm1, %mm3
movq %mm2, -16(rp)
movq %mm3, -8(rp)
L(ast): movq (up), %mm2
psrlq %mm4, %mm2
movq %mm2, (rp)
emms
FUNC_EXIT()
ret
EPILOGUE()
|
