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
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
|
dnl S/390-32 mpn_sqr_basecase.
dnl Copyright 2011 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 z900 ?
C z990 23
C z9 ?
C z10 ?
C z196 ?
C TODO
C * Clean up.
C * Stop iterating addmul_1 loop at latest for n = 2, implement longer tail.
C This will ask for basecase handling of n = 3.
C * Update counters and pointers more straightforwardly, possibly lowering
C register usage.
C * Should we use this allocation-free style for more sqr_basecase asm
C implementations? The only disadvantage is that it requires R != U.
C * Replace loops by faster code. The mul_1 and addmul_1 loops could be sped
C up by about 10%. The sqr_diag_addlsh1 loop could probably be sped up even
C more.
C INPUT PARAMETERS
define(`rp', `%r2')
define(`up', `%r3')
define(`n', `%r4')
define(`zero', `%r8')
define(`rp_saved', `%r9')
define(`up_saved', `%r13')
define(`n_saved', `%r14')
ASM_START()
PROLOGUE(mpn_sqr_basecase)
ahi n, -2
jhe L(ge2)
C n = 1
l %r5, 0(up)
mlr %r4, %r5
st %r5, 0(rp)
st %r4, 4(rp)
br %r14
L(ge2): jne L(gen)
C n = 2
stm %r6, %r8, 24(%r15)
lhi zero, 0
l %r5, 0(up)
mlr %r4, %r5 C u0 * u0
l %r1, 4(up)
mlr %r0, %r1 C u1 * u1
st %r5, 0(rp)
l %r7, 0(up)
ml %r6, 4(up) C u0 * u1
alr %r7, %r7
alcr %r6, %r6
alcr %r0, zero
alr %r4, %r7
alcr %r1, %r6
alcr %r0, zero
st %r4, 4(rp)
st %r1, 8(rp)
st %r0, 12(rp)
lm %r6, %r8, 24(%r15)
br %r14
L(gen):
C mul_1 =======================================================================
stm %r6, %r14, 24(%r15)
lhi zero, 0
lr up_saved, up
lr rp_saved, rp
lr n_saved, n
l %r6, 0(up)
l %r11, 4(up)
lhi %r12, 8 C init index register
mlr %r10, %r6
lr %r5, n
st %r11, 4(rp)
cr %r15, %r15 C clear carry flag
L(tm): l %r1, 0(%r12,up)
mlr %r0, %r6
alcr %r1, %r10
lr %r10, %r0 C copy high part to carry limb
st %r1, 0(%r12,rp)
la %r12, 4(%r12)
brct %r5, L(tm)
alcr %r0, zero
st %r0, 0(%r12,rp)
C addmul_1 loop ===============================================================
ahi n, -1
je L(outer_end)
L(outer_loop):
la rp, 8(rp) C rp += 2
la up, 4(up) C up += 1
l %r6, 0(up)
l %r11, 4(up)
lhi %r12, 8 C init index register
mlr %r10, %r6
lr %r5, n
al %r11, 4(rp)
st %r11, 4(rp)
L(tam): l %r1, 0(%r12,up)
l %r7, 0(%r12,rp)
mlr %r0, %r6
alcr %r1, %r7
alcr %r0, zero
alr %r1, %r10
lr %r10, %r0
st %r1, 0(%r12,rp)
la %r12, 4(%r12)
brct %r5, L(tam)
alcr %r0, zero
st %r0, 0(%r12,rp)
brct n, L(outer_loop)
L(outer_end):
l %r6, 4(up)
l %r1, 8(up)
lr %r7, %r0 C Same as: l %r7, 12(,rp)
mlr %r0, %r6
alr %r1, %r7
alcr %r0, zero
st %r1, 12(rp)
st %r0, 16(rp)
C sqr_diag_addlsh1 ============================================================
define(`up', `up_saved')
define(`rp', `rp_saved')
la n, 1(n_saved)
l %r1, 0(up)
mlr %r0, %r1
st %r1, 0(rp)
C clr %r15, %r15 C clear carry (already clear per above)
L(top): l %r11, 4(up)
la up, 4(up)
l %r6, 4(rp)
l %r7, 8(rp)
mlr %r10, %r11
alcr %r6, %r6
alcr %r7, %r7
alcr %r10, zero C propagate carry to high product limb
alr %r6, %r0
alcr %r7, %r11
stm %r6, %r7, 4(rp)
la rp, 8(rp)
lr %r0, %r10 C copy carry limb
brct n, L(top)
alcr %r0, zero
st %r0, 4(rp)
lm %r6, %r14, 24(%r15)
br %r14
EPILOGUE()
|
