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
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
|
dnl Alpha mpn_bdiv_dbm1c.
dnl Copyright 2008 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 EV4: 42
C EV5: 18
C EV6: 3
C TODO
C * Try less unrolling, 2-way should give the same performance.
C * Optimize feed-in and wind-down code, for speed, and perhaps further for
C code size.
C * This runs optimally given the algorithm, r8 is on a 3 operation recurrency
C path. We have not tried very hard to find a better algorithm. Perhaps
C it would be a good task for the GNU superoptimizer.
C INPUT PARAMETERS
define(`rp', `r16')
define(`up', `r17')
define(`n', `r18')
define(`bd', `r19')
define(`cy', `r19')
ASM_START()
PROLOGUE(mpn_bdiv_dbm1c)
mov r20, r8
ldq r24, 0(r17)
and r18, 3, r28
lda r18, -4(r18)
beq r28, L(b0)
cmpeq r28, 1, r21
bne r21, L(b1)
cmpeq r28, 2, r21
bne r21, L(b2)
L(b3): ldq r2, 8(r17)
ldq r3, 16(r17)
bgt r18, L(gt3)
mulq r24, r19, r5 C U1
umulh r24, r19, r21 C U1
mulq r2, r19, r6 C U1
umulh r2, r19, r22 C U1
mulq r3, r19, r7 C U1
umulh r3, r19, r23 C U1
lda r16, -32(r16)
br L(cj3)
L(gt3): ldq r0, 24(r17)
mulq r24, r19, r5 C U1
umulh r24, r19, r21 C U1
ldq r1, 32(r17)
mulq r2, r19, r6 C U1
umulh r2, r19, r22 C U1
ldq r2, 40(r17)
mulq r3, r19, r7 C U1
umulh r3, r19, r23 C U1
ldq r3, 48(r17)
lda r18, -4(r18)
lda r17, 56(r17)
mulq r0, r19, r4 C U1
bgt r18, L(L3)
br L(cj7)
L(b2): ldq r3, 8(r17)
bgt r18, L(gt2)
mulq r24, r19, r6 C U1
umulh r24, r19, r22 C U1
mulq r3, r19, r7 C U1
umulh r3, r19, r23 C U1
lda r16, -40(r16)
br L(cj2)
L(gt2): ldq r0, 16(r17)
ldq r1, 24(r17)
mulq r24, r19, r6 C U1
umulh r24, r19, r22 C U1
ldq r2, 32(r17)
mulq r3, r19, r7 C U1
umulh r3, r19, r23 C U1
ldq r3, 40(r17)
lda r18, -4(r18)
lda r17, 48(r17)
mulq r0, r19, r4 C U1
umulh r0, r19, r20 C U1
lda r16, -8(r16)
bgt r18, L(gt6)
mulq r1, r19, r5 C U1
br L(cj6)
L(gt6): ldq r0, 0(r17)
mulq r1, r19, r5 C U1
br L(L2)
L(b1): bgt r18, L(gt1)
mulq r24, r19, r7 C U1
umulh r24, r19, r23 C U1
lda r16, -48(r16)
br L(cj1)
L(gt1): ldq r0, 8(r17)
ldq r1, 16(r17)
ldq r2, 24(r17)
mulq r24, r19, r7 C U1
umulh r24, r19, r23 C U1
ldq r3, 32(r17)
lda r18, -4(r18)
lda r17, 40(r17)
mulq r0, r19, r4 C U1
umulh r0, r19, r20 C U1
lda r16, -16(r16)
bgt r18, L(gt5)
mulq r1, r19, r5 C U1
umulh r1, r19, r21 C U1
mulq r2, r19, r6 C U1
br L(cj5)
L(gt5): ldq r0, 0(r17)
mulq r1, r19, r5 C U1
umulh r1, r19, r21 C U1
ldq r1, 8(r17)
mulq r2, r19, r6 C U1
br L(L1)
L(b0): ldq r1, 8(r17)
ldq r2, 16(r17)
ldq r3, 24(r17)
lda r17, 32(r17)
lda r16, -24(r16)
mulq r24, r19, r4 C U1
umulh r24, r19, r20 C U1
bgt r18, L(gt4)
mulq r1, r19, r5 C U1
umulh r1, r19, r21 C U1
mulq r2, r19, r6 C U1
umulh r2, r19, r22 C U1
mulq r3, r19, r7 C U1
br L(cj4)
L(gt4): ldq r0, 0(r17)
mulq r1, r19, r5 C U1
umulh r1, r19, r21 C U1
ldq r1, 8(r17)
mulq r2, r19, r6 C U1
umulh r2, r19, r22 C U1
ldq r2, 16(r17)
mulq r3, r19, r7 C U1
br L(L0)
C *** MAIN LOOP START ***
ALIGN(16)
L(top): mulq r0, r19, r4 C U1
subq r8, r28, r8
L(L3): umulh r0, r19, r20 C U1
cmpult r8, r5, r28
ldq r0, 0(r17)
subq r8, r5, r8
addq r21, r28, r28
stq r8, 0(r16)
mulq r1, r19, r5 C U1
subq r8, r28, r8
L(L2): umulh r1, r19, r21 C U1
cmpult r8, r6, r28
ldq r1, 8(r17)
subq r8, r6, r8
addq r22, r28, r28
stq r8, 8(r16)
mulq r2, r19, r6 C U1
subq r8, r28, r8
L(L1): umulh r2, r19, r22 C U1
cmpult r8, r7, r28
ldq r2, 16(r17)
subq r8, r7, r8
addq r23, r28, r28
stq r8, 16(r16)
mulq r3, r19, r7 C U1
subq r8, r28, r8
L(L0): umulh r3, r19, r23 C U1
cmpult r8, r4, r28
ldq r3, 24(r17)
subq r8, r4, r8
addq r20, r28, r28
stq r8, 24(r16)
lda r18, -4(r18)
lda r17, 32(r17)
lda r16, 32(r16)
bgt r18, L(top)
C *** MAIN LOOP END ***
mulq r0, r19, r4 C U1
subq r8, r28, r8
L(cj7): umulh r0, r19, r20 C U1
cmpult r8, r5, r28
subq r8, r5, r8
addq r21, r28, r28
stq r8, 0(r16)
mulq r1, r19, r5 C U1
subq r8, r28, r8
L(cj6): umulh r1, r19, r21 C U1
cmpult r8, r6, r28
subq r8, r6, r8
addq r22, r28, r28
stq r8, 8(r16)
mulq r2, r19, r6 C U1
subq r8, r28, r8
L(cj5): umulh r2, r19, r22 C U1
cmpult r8, r7, r28
subq r8, r7, r8
addq r23, r28, r28
stq r8, 16(r16)
mulq r3, r19, r7 C U1
subq r8, r28, r8
L(cj4): umulh r3, r19, r23 C U1
cmpult r8, r4, r28
subq r8, r4, r8
addq r20, r28, r28
stq r8, 24(r16)
subq r8, r28, r8
L(cj3): cmpult r8, r5, r28
subq r8, r5, r8
addq r21, r28, r28
stq r8, 32(r16)
subq r8, r28, r8
L(cj2): cmpult r8, r6, r28
subq r8, r6, r8
addq r22, r28, r28
stq r8, 40(r16)
subq r8, r28, r8
L(cj1): cmpult r8, r7, r28
subq r8, r7, r8
addq r23, r28, r28
stq r8, 48(r16)
subq r8, r28, r0
ret r31, (r26), 1
EPILOGUE()
ASM_END()
|
