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
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
|
dnl IA-64 mpn_lshift/mpn_rshift.
dnl Copyright 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation,
dnl 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 Itanium: 2.0
C Itanium 2: 1.0
C This code is scheduled deeply since the plain shift instructions shr and shl
C have a latency of 4 (on Itanium) or 3 (on Itanium 2). Poor scheduling of
C these instructions cause a 10 cycle replay trap on Itanium.
C TODO
C * Optimize function entry and feed-in code.
C INPUT PARAMETERS
define(`rp',`r32')
define(`up',`r33')
define(`n',`r34')
define(`cnt',`r35')
define(`tnc',`r9')
ifdef(`OPERATION_lshift',`
define(`FSH',`shl')
define(`BSH',`shr.u')
define(`UPD',`-8')
define(`POFF',`-512')
define(`PUPD',`-32')
define(`func',`mpn_lshift')
')
ifdef(`OPERATION_rshift',`
define(`FSH',`shr.u')
define(`BSH',`shl')
define(`UPD',`8')
define(`POFF',`512')
define(`PUPD',`32')
define(`func',`mpn_rshift')
')
MULFUNC_PROLOGUE(mpn_lshift mpn_rshift)
ASM_START()
PROLOGUE(func)
.prologue
.save ar.lc, r2
.body
ifdef(`HAVE_ABI_32',
` addp4 rp = 0, rp C M I
addp4 up = 0, up C M I
sxt4 n = n C M I
zxt4 cnt = cnt C I
;;
')
{.mmi; cmp.lt p14, p15 = 4, n C M I
and r14 = 3, n C M I
mov.i r2 = ar.lc C I0
}{.mmi; add r15 = -1, n C M I
sub tnc = 64, cnt C M I
add r16 = -5, n
;;
}{.mmi; cmp.eq p6, p0 = 1, r14 C M I
cmp.eq p7, p0 = 2, r14 C M I
shr.u n = r16, 2 C I0
}{.mmi; cmp.eq p8, p0 = 3, r14 C M I
ifdef(`OPERATION_lshift',
` shladd up = r15, 3, up C M I
shladd rp = r15, 3, rp') C M I
;;
}{.mmi; add r11 = POFF, up C M I
ld8 r10 = [up], UPD C M01
mov.i ar.lc = n C I0
}{.bbb;
(p6) br.dptk .Lb01
(p7) br.dptk .Lb10
(p8) br.dptk .Lb11
;;
}
.Lb00: ld8 r19 = [up], UPD
;;
ld8 r16 = [up], UPD
;;
ld8 r17 = [up], UPD
BSH r8 = r10, tnc C function return value
(p14) br.cond.dptk .grt4
FSH r24 = r10, cnt
BSH r25 = r19, tnc
;;
FSH r26 = r19, cnt
BSH r27 = r16, tnc
;;
FSH r20 = r16, cnt
BSH r21 = r17, tnc
;;
or r14 = r25, r24
FSH r22 = r17, cnt
BSH r23 = r10, tnc
br .Lr4
.grt4: FSH r24 = r10, cnt
BSH r25 = r19, tnc
;;
ld8 r18 = [up], UPD
FSH r26 = r19, cnt
BSH r27 = r16, tnc
;;
ld8 r19 = [up], UPD
FSH r20 = r16, cnt
BSH r21 = r17, tnc
;;
ld8 r16 = [up], UPD
FSH r22 = r17, cnt
BSH r23 = r18, tnc
;;
or r14 = r25, r24
ld8 r17 = [up], UPD
br.cloop.dpnt .Ltop
br .Lbot
.Lb01:
(p15) BSH r8 = r10, tnc C function return value I
(p15) FSH r22 = r10, cnt C I
(p15) br.cond.dptk .Lr1 C return B
.grt1: ld8 r18 = [up], UPD
;;
ld8 r19 = [up], UPD
BSH r8 = r10, tnc C function return value
;;
ld8 r16 = [up], UPD
FSH r22 = r10, cnt
BSH r23 = r18, tnc
;;
ld8 r17 = [up], UPD
br.cloop.dpnt .grt5
;;
FSH r24 = r18, cnt
BSH r25 = r19, tnc
;;
or r15 = r23, r22
FSH r26 = r19, cnt
BSH r27 = r16, tnc
;;
FSH r20 = r16, cnt
BSH r21 = r17, tnc
br .Lr5
.grt5: FSH r24 = r18, cnt
BSH r25 = r19, tnc
;;
ld8 r18 = [up], UPD
FSH r26 = r19, cnt
BSH r27 = r16, tnc
;;
ld8 r19 = [up], UPD
FSH r20 = r16, cnt
BSH r21 = r17, tnc
;;
or r15 = r23, r22
ld8 r16 = [up], UPD
br .LL01
.Lb10: ld8 r17 = [up], UPD
(p14) br.cond.dptk .grt2
BSH r8 = r10, tnc C function return value
;;
FSH r20 = r10, cnt
BSH r21 = r17, tnc
;;
or r14 = r21, r20
FSH r22 = r17, cnt
br .Lr2 C return
.grt2: ld8 r18 = [up], UPD
BSH r8 = r10, tnc C function return value
;;
ld8 r19 = [up], UPD
FSH r20 = r10, cnt
BSH r21 = r17, tnc
;;
ld8 r16 = [up], UPD
FSH r22 = r17, cnt
BSH r23 = r18, tnc
;;
ld8 r17 = [up], UPD
br.cloop.dpnt .grt6
;;
or r14 = r21, r20
FSH r24 = r18, cnt
BSH r25 = r19, tnc
;;
FSH r26 = r19, cnt
BSH r27 = r16, tnc
br .Lr6
.grt6: or r14 = r21, r20
FSH r24 = r18, cnt
BSH r25 = r19, tnc
;;
ld8 r18 = [up], UPD
FSH r26 = r19, cnt
BSH r27 = r16, tnc
;;
ld8 r19 = [up], UPD
br .LL10
.Lb11: ld8 r16 = [up], UPD
;;
ld8 r17 = [up], UPD
BSH r8 = r10, tnc C function return value
(p14) br.cond.dptk .grt3
;;
FSH r26 = r10, cnt
BSH r27 = r16, tnc
;;
FSH r20 = r16, cnt
BSH r21 = r17, tnc
;;
or r15 = r27, r26
FSH r22 = r17, cnt
br .Lr3 C return
.grt3: ld8 r18 = [up], UPD
FSH r26 = r10, cnt
BSH r27 = r16, tnc
;;
ld8 r19 = [up], UPD
FSH r20 = r16, cnt
BSH r21 = r17, tnc
;;
ld8 r16 = [up], UPD
FSH r22 = r17, cnt
BSH r23 = r18, tnc
;;
ld8 r17 = [up], UPD
br.cloop.dpnt .grt7
or r15 = r27, r26
FSH r24 = r18, cnt
BSH r25 = r19, tnc
br .Lr7
.grt7: or r15 = r27, r26
FSH r24 = r18, cnt
BSH r25 = r19, tnc
ld8 r18 = [up], UPD
br .LL11
C *** MAIN LOOP START ***
ALIGN(32)
.Ltop:
{.mmi; st8 [rp] = r14, UPD C M2
or r15 = r27, r26 C M3
FSH r24 = r18, cnt C I0
}{.mmi; ld8 r18 = [up], UPD C M1
lfetch [r11], PUPD
BSH r25 = r19, tnc C I1
;; }
.LL11:
{.mmi; st8 [rp] = r15, UPD
or r14 = r21, r20
FSH r26 = r19, cnt
}{.mmi; ld8 r19 = [up], UPD
nop.m 0
BSH r27 = r16, tnc
;; }
.LL10:
{.mmi; st8 [rp] = r14, UPD
or r15 = r23, r22
FSH r20 = r16, cnt
}{.mmi; ld8 r16 = [up], UPD
nop.m 0
BSH r21 = r17, tnc
;; }
.LL01:
{.mmi; st8 [rp] = r15, UPD
or r14 = r25, r24
FSH r22 = r17, cnt
}{.mib; ld8 r17 = [up], UPD
BSH r23 = r18, tnc
br.cloop.dptk .Ltop
;; }
C *** MAIN LOOP END ***
.Lbot: or r15 = r27, r26
FSH r24 = r18, cnt
BSH r25 = r19, tnc
st8 [rp] = r14, UPD
;;
.Lr7: or r14 = r21, r20
FSH r26 = r19, cnt
BSH r27 = r16, tnc
st8 [rp] = r15, UPD
;;
.Lr6: or r15 = r23, r22
FSH r20 = r16, cnt
BSH r21 = r17, tnc
st8 [rp] = r14, UPD
;;
.Lr5: st8 [rp] = r15, UPD
or r14 = r25, r24
FSH r22 = r17, cnt
;;
.Lr4: or r15 = r27, r26
st8 [rp] = r14, UPD
;;
.Lr3: or r14 = r21, r20
st8 [rp] = r15, UPD
;;
.Lr2: st8 [rp] = r14, UPD
;;
.Lr1: st8 [rp] = r22, UPD C M23
mov ar.lc = r2 C I0
br.ret.sptk.many b0 C B
EPILOGUE(func)
ASM_END()
|
