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
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
|
dnl Alpha mpn_divexact_by3c -- mpn division by 3, expecting no remainder.
dnl Copyright 2004, 2005 Free Software Foundation, Inc.
dnl This file is part of the GNU MP Library.
dnl The GNU MP Library is free software; you can redistribute it and/or modify
dnl it under the terms of the GNU Lesser General Public License as published
dnl by the Free Software Foundation; either version 3 of the License, or (at
dnl your option) any later version.
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 Lesser General Public
dnl License for more details.
dnl You should have received a copy of the GNU Lesser General Public License
dnl along with the GNU MP Library. If not, see http://www.gnu.org/licenses/.
include(`../config.m4')
C cycles/limb
C EV4: 22
C EV5: 11.5
C EV6: 6.3
C TODO
C * Trim this to 6.0 c/l for ev6.
C * Write special ev5 version, should reach 9 c/l, and could be smaller.
C * Try prefetch for destination, using lds.
C * Improve feed-in code, by moving initial mulq earlier; make initial load
C to u0/u0 to save some copying.
C * Combine u0 and u2, u1 and u3.
C INPUT PARAMETERS
define(`rp', `r16')
define(`up', `r17')
define(`n', `r18')
define(`cy', `r19')
ASM_START()
DATASTART(L(LC))
.quad 0xAAAAAAAAAAAAAAAB
.quad 0x5555555555555555
.quad 0xAAAAAAAAAAAAAAAA
DATAEND()
define(`xAAAAAAAAAAAAAAAB', `r20')
define(`x5555555555555555', `r21')
define(`xAAAAAAAAAAAAAAAA', `r22')
define(`u0', `r0') define(`u1', `r1')
define(`u2', `r2') define(`u3', `r3')
define(`l0', `r25') define(`x', `r8')
define(`q0', `r4') define(`q1', `r5')
define(`p6', `r6') define(`p7', `r7')
define(`t0', `r23') define(`t1', `r24')
define(`cymask',`r28')
PROLOGUE(mpn_divexact_by3c,gp)
ldq r28, 0(up) C load first limb early
C Put magic constants in registers
lda r0, L(LC)
ldq xAAAAAAAAAAAAAAAB, 0(r0)
ldq x5555555555555555, 8(r0)
ldq xAAAAAAAAAAAAAAAA, 16(r0)
C Compute initial l0 value
cmpeq cy, 1, p6
cmpeq cy, 2, p7
negq p6, p6
and p6, x5555555555555555, l0
cmovne p7, xAAAAAAAAAAAAAAAA, l0
C Feed-in depending on (n mod 4)
and n, 3, r8
lda n, -3(n)
cmpeq r8, 1, r4
cmpeq r8, 2, r5
bne r4, $Lb01
bne r5, $Lb10
beq r8, $Lb00
$Lb11: ldq u3, 8(up)
lda up, -24(up)
lda rp, -24(rp)
mulq r28, xAAAAAAAAAAAAAAAB, q0
mov r28, u2
br r31, $L11
$Lb00: ldq u2, 8(up)
lda up, -16(up)
lda rp, -16(rp)
mulq r28, xAAAAAAAAAAAAAAAB, q1
mov r28, u1
br r31, $L00
$Lb01: lda rp, -8(rp)
mulq r28, xAAAAAAAAAAAAAAAB, q0
mov r28, u0
blt n, $Lcj1
ldq u1, 8(up)
lda up, -8(up)
br r31, $L01
$Lb10: ldq u0, 8(up)
mulq r28, xAAAAAAAAAAAAAAAB, q1
mov r28, u3
blt n, $Lend
ALIGN(16)
$Ltop:
C 0
cmpult u3, cy, cy C L0
mulq u0, xAAAAAAAAAAAAAAAB, q0 C U1
ldq u1, 16(up) C L1
addq q1, l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
unop
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, 0(rp) C L1
unop
$L01:
C 0
cmpult u0, cy, cy C L0
mulq u1, xAAAAAAAAAAAAAAAB, q1 C U1
ldq u2, 24(up) C L1
addq q0, l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
unop
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, 8(rp) C L1
unop
$L00:
C 0
cmpult u1, cy, cy C L0
mulq u2, xAAAAAAAAAAAAAAAB, q0 C U1
ldq u3, 32(up) C L1
addq q1, l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
unop
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, 16(rp) C L1
unop
$L11:
C 0
cmpult u2, cy, cy C L0
mulq u3, xAAAAAAAAAAAAAAAB, q1 C U1
ldq u0, 40(up) C L1
addq q0, l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
lda n, -4(n) C L1 bookkeeping
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, 24(rp) C L1
lda up, 32(up)
C
ldl r31, 256(up) C prefetch
unop
lda rp, 32(rp)
bge n, $Ltop C U1
C *** MAIN LOOP END ***
$Lend:
cmpult u3, cy, cy C L0
mulq u0, xAAAAAAAAAAAAAAAB, q0 C U1
unop
addq q1, l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
unop
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, 0(rp) C L1
unop
$Lcj1:
cmpult u0, cy, cy C L0
addq q0, l0, x C U0
cmpult x5555555555555555, x, p6 C U0
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
addq p6, cy, cy
addq p7, cy, r0
stq x, 8(rp) C L1
ret r31,(r26),1
EPILOGUE()
ASM_END()
C This is useful for playing with various schedules.
C Expand as: one(0)one(1)one(2)one(3)
define(`one',`
C 0
cmpult `$'eval(($1+3)%4), cy, cy C L0
mulq `$'$1, xAAAAAAAAAAAAAAAB, `$'eval(4+$1%2) C U1
ldq `$'eval(($1+1)%4), eval($1*8+16)(up) C L1
addq `$'eval(4+($1+1)%2), l0, x C U0
C 1
negq cy, cymask C L0
unop C U1
unop C L1
cmpult x5555555555555555, x, p6 C U0
C 2
cmpult xAAAAAAAAAAAAAAAA, x, p7 C U1
unop
unop
negq p6, t0 C L0
C 3
negq p7, t1 C L0
and cymask, x5555555555555555, l0 C U1
addq p6, cy, cy
and t0, x5555555555555555, t0
C 4
and t1, x5555555555555555, t1
addq p7, cy, cy
unop
addq t0, l0, l0
C 5
addq t1, l0, l0
unop
stq x, eval($1*8)(rp) C L1
unop
')
|
