-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy patharith_riscv64.s
457 lines (448 loc) · 9.03 KB
/
arith_riscv64.s
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
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
// Copyright 2025 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Code generated by 'go generate' (with ./internal/asmgen). DO NOT EDIT.
//go:build !math_big_pure_go
#include "textflag.h"
// func addVV(z, x, y []Word) (c Word)
TEXT ·addVV(SB), NOSPLIT, $0
MOV z_len+8(FP), X5
MOV x_base+24(FP), X6
MOV y_base+48(FP), X7
MOV z_base+0(FP), X8
// compute unrolled loop lengths
AND $3, X5, X9
SRL $2, X5
XOR X28, X28 // clear carry
loop1:
BEQZ X9, loop1done
loop1cont:
// unroll 1X
MOV 0(X6), X10
MOV 0(X7), X11
ADD X11, X10 // ADCS X11, X10, X10 (cr=X28)
SLTU X11, X10, X31 // ...
ADD X28, X10 // ...
SLTU X28, X10, X28 // ...
ADD X31, X28 // ...
MOV X10, 0(X8)
ADD $8, X6
ADD $8, X7
ADD $8, X8
SUB $1, X9
BNEZ X9, loop1cont
loop1done:
loop4:
BEQZ X5, loop4done
loop4cont:
// unroll 4X
MOV 0(X6), X9
MOV 8(X6), X10
MOV 16(X6), X11
MOV 24(X6), X12
MOV 0(X7), X13
MOV 8(X7), X14
MOV 16(X7), X15
MOV 24(X7), X16
ADD X13, X9 // ADCS X13, X9, X9 (cr=X28)
SLTU X13, X9, X31 // ...
ADD X28, X9 // ...
SLTU X28, X9, X28 // ...
ADD X31, X28 // ...
ADD X14, X10 // ADCS X14, X10, X10 (cr=X28)
SLTU X14, X10, X31 // ...
ADD X28, X10 // ...
SLTU X28, X10, X28 // ...
ADD X31, X28 // ...
ADD X15, X11 // ADCS X15, X11, X11 (cr=X28)
SLTU X15, X11, X31 // ...
ADD X28, X11 // ...
SLTU X28, X11, X28 // ...
ADD X31, X28 // ...
ADD X16, X12 // ADCS X16, X12, X12 (cr=X28)
SLTU X16, X12, X31 // ...
ADD X28, X12 // ...
SLTU X28, X12, X28 // ...
ADD X31, X28 // ...
MOV X9, 0(X8)
MOV X10, 8(X8)
MOV X11, 16(X8)
MOV X12, 24(X8)
ADD $32, X6
ADD $32, X7
ADD $32, X8
SUB $1, X5
BNEZ X5, loop4cont
loop4done:
MOV X28, c+72(FP)
RET
// func subVV(z, x, y []Word) (c Word)
TEXT ·subVV(SB), NOSPLIT, $0
MOV z_len+8(FP), X5
MOV x_base+24(FP), X6
MOV y_base+48(FP), X7
MOV z_base+0(FP), X8
// compute unrolled loop lengths
AND $3, X5, X9
SRL $2, X5
XOR X28, X28 // clear carry
loop1:
BEQZ X9, loop1done
loop1cont:
// unroll 1X
MOV 0(X6), X10
MOV 0(X7), X11
SLTU X28, X10, X31 // SBCS X11, X10, X10
SUB X28, X10 // ...
SLTU X11, X10, X28 // ...
SUB X11, X10 // ...
ADD X31, X28 // ...
MOV X10, 0(X8)
ADD $8, X6
ADD $8, X7
ADD $8, X8
SUB $1, X9
BNEZ X9, loop1cont
loop1done:
loop4:
BEQZ X5, loop4done
loop4cont:
// unroll 4X
MOV 0(X6), X9
MOV 8(X6), X10
MOV 16(X6), X11
MOV 24(X6), X12
MOV 0(X7), X13
MOV 8(X7), X14
MOV 16(X7), X15
MOV 24(X7), X16
SLTU X28, X9, X31 // SBCS X13, X9, X9
SUB X28, X9 // ...
SLTU X13, X9, X28 // ...
SUB X13, X9 // ...
ADD X31, X28 // ...
SLTU X28, X10, X31 // SBCS X14, X10, X10
SUB X28, X10 // ...
SLTU X14, X10, X28 // ...
SUB X14, X10 // ...
ADD X31, X28 // ...
SLTU X28, X11, X31 // SBCS X15, X11, X11
SUB X28, X11 // ...
SLTU X15, X11, X28 // ...
SUB X15, X11 // ...
ADD X31, X28 // ...
SLTU X28, X12, X31 // SBCS X16, X12, X12
SUB X28, X12 // ...
SLTU X16, X12, X28 // ...
SUB X16, X12 // ...
ADD X31, X28 // ...
MOV X9, 0(X8)
MOV X10, 8(X8)
MOV X11, 16(X8)
MOV X12, 24(X8)
ADD $32, X6
ADD $32, X7
ADD $32, X8
SUB $1, X5
BNEZ X5, loop4cont
loop4done:
MOV X28, c+72(FP)
RET
// func lshVU(z, x []Word, s uint) (c Word)
TEXT ·lshVU(SB), NOSPLIT, $0
MOV z_len+8(FP), X5
BEQZ X5, ret0
MOV s+48(FP), X6
MOV x_base+24(FP), X7
MOV z_base+0(FP), X8
// run loop backward
SLL $3, X5, X9
ADD X9, X7
SLL $3, X5, X9
ADD X9, X8
// shift first word into carry
MOV -8(X7), X9
MOV $64, X10
SUB X6, X10
SRL X10, X9, X11
SLL X6, X9
MOV X11, c+56(FP)
// shift remaining words
SUB $1, X5
// compute unrolled loop lengths
AND $3, X5, X11
SRL $2, X5
loop1:
BEQZ X11, loop1done
loop1cont:
// unroll 1X
MOV -16(X7), X12
SRL X10, X12, X13
OR X9, X13
SLL X6, X12, X9
MOV X13, -8(X8)
ADD $-8, X7
ADD $-8, X8
SUB $1, X11
BNEZ X11, loop1cont
loop1done:
loop4:
BEQZ X5, loop4done
loop4cont:
// unroll 4X
MOV -16(X7), X11
MOV -24(X7), X12
MOV -32(X7), X13
MOV -40(X7), X14
SRL X10, X11, X15
OR X9, X15
SLL X6, X11, X9
SRL X10, X12, X11
OR X9, X11
SLL X6, X12, X9
SRL X10, X13, X12
OR X9, X12
SLL X6, X13, X9
SRL X10, X14, X13
OR X9, X13
SLL X6, X14, X9
MOV X15, -8(X8)
MOV X11, -16(X8)
MOV X12, -24(X8)
MOV X13, -32(X8)
ADD $-32, X7
ADD $-32, X8
SUB $1, X5
BNEZ X5, loop4cont
loop4done:
// store final shifted bits
MOV X9, -8(X8)
RET
ret0:
MOV X0, c+56(FP)
RET
// func rshVU(z, x []Word, s uint) (c Word)
TEXT ·rshVU(SB), NOSPLIT, $0
MOV z_len+8(FP), X5
BEQZ X5, ret0
MOV s+48(FP), X6
MOV x_base+24(FP), X7
MOV z_base+0(FP), X8
// shift first word into carry
MOV 0(X7), X9
MOV $64, X10
SUB X6, X10
SLL X10, X9, X11
SRL X6, X9
MOV X11, c+56(FP)
// shift remaining words
SUB $1, X5
// compute unrolled loop lengths
AND $3, X5, X11
SRL $2, X5
loop1:
BEQZ X11, loop1done
loop1cont:
// unroll 1X
MOV 8(X7), X12
SLL X10, X12, X13
OR X9, X13
SRL X6, X12, X9
MOV X13, 0(X8)
ADD $8, X7
ADD $8, X8
SUB $1, X11
BNEZ X11, loop1cont
loop1done:
loop4:
BEQZ X5, loop4done
loop4cont:
// unroll 4X
MOV 8(X7), X11
MOV 16(X7), X12
MOV 24(X7), X13
MOV 32(X7), X14
SLL X10, X11, X15
OR X9, X15
SRL X6, X11, X9
SLL X10, X12, X11
OR X9, X11
SRL X6, X12, X9
SLL X10, X13, X12
OR X9, X12
SRL X6, X13, X9
SLL X10, X14, X13
OR X9, X13
SRL X6, X14, X9
MOV X15, 0(X8)
MOV X11, 8(X8)
MOV X12, 16(X8)
MOV X13, 24(X8)
ADD $32, X7
ADD $32, X8
SUB $1, X5
BNEZ X5, loop4cont
loop4done:
// store final shifted bits
MOV X9, 0(X8)
RET
ret0:
MOV X0, c+56(FP)
RET
// func mulAddVWW(z, x []Word, m, a Word) (c Word)
TEXT ·mulAddVWW(SB), NOSPLIT, $0
MOV m+48(FP), X5
MOV a+56(FP), X6
MOV z_len+8(FP), X7
MOV x_base+24(FP), X8
MOV z_base+0(FP), X9
// compute unrolled loop lengths
AND $3, X7, X10
SRL $2, X7
loop1:
BEQZ X10, loop1done
loop1cont:
// unroll 1X
MOV 0(X8), X11
// synthetic carry, one column at a time
MUL X5, X11, X12
MULHU X5, X11, X13
ADD X6, X12, X11 // ADDS X6, X12, X11 (cr=X28)
SLTU X6, X11, X28 // ...
ADD X28, X13, X6 // ADC $0, X13, X6
MOV X11, 0(X9)
ADD $8, X8
ADD $8, X9
SUB $1, X10
BNEZ X10, loop1cont
loop1done:
loop4:
BEQZ X7, loop4done
loop4cont:
// unroll 4X
MOV 0(X8), X10
MOV 8(X8), X11
MOV 16(X8), X12
MOV 24(X8), X13
// synthetic carry, one column at a time
MUL X5, X10, X14
MULHU X5, X10, X15
ADD X6, X14, X10 // ADDS X6, X14, X10 (cr=X28)
SLTU X6, X10, X28 // ...
ADD X28, X15, X6 // ADC $0, X15, X6
MUL X5, X11, X14
MULHU X5, X11, X15
ADD X6, X14, X11 // ADDS X6, X14, X11 (cr=X28)
SLTU X6, X11, X28 // ...
ADD X28, X15, X6 // ADC $0, X15, X6
MUL X5, X12, X14
MULHU X5, X12, X15
ADD X6, X14, X12 // ADDS X6, X14, X12 (cr=X28)
SLTU X6, X12, X28 // ...
ADD X28, X15, X6 // ADC $0, X15, X6
MUL X5, X13, X14
MULHU X5, X13, X15
ADD X6, X14, X13 // ADDS X6, X14, X13 (cr=X28)
SLTU X6, X13, X28 // ...
ADD X28, X15, X6 // ADC $0, X15, X6
MOV X10, 0(X9)
MOV X11, 8(X9)
MOV X12, 16(X9)
MOV X13, 24(X9)
ADD $32, X8
ADD $32, X9
SUB $1, X7
BNEZ X7, loop4cont
loop4done:
MOV X6, c+64(FP)
RET
// func addMulVVWW(z, x, y []Word, m, a Word) (c Word)
TEXT ·addMulVVWW(SB), NOSPLIT, $0
MOV m+72(FP), X5
MOV a+80(FP), X6
MOV z_len+8(FP), X7
MOV x_base+24(FP), X8
MOV y_base+48(FP), X9
MOV z_base+0(FP), X10
// compute unrolled loop lengths
AND $3, X7, X11
SRL $2, X7
loop1:
BEQZ X11, loop1done
loop1cont:
// unroll 1X
MOV 0(X8), X12
MOV 0(X9), X13
// synthetic carry, one column at a time
MUL X5, X13, X14
MULHU X5, X13, X15
ADD X12, X14 // ADDS X12, X14, X14 (cr=X28)
SLTU X12, X14, X28 // ...
ADD X28, X15 // ADC $0, X15, X15
ADD X6, X14, X13 // ADDS X6, X14, X13 (cr=X28)
SLTU X6, X13, X28 // ...
ADD X28, X15, X6 // ADC $0, X15, X6
MOV X13, 0(X10)
ADD $8, X8
ADD $8, X9
ADD $8, X10
SUB $1, X11
BNEZ X11, loop1cont
loop1done:
loop4:
BEQZ X7, loop4done
loop4cont:
// unroll 4X
MOV 0(X8), X11
MOV 8(X8), X12
MOV 16(X8), X13
MOV 24(X8), X14
MOV 0(X9), X15
MOV 8(X9), X16
MOV 16(X9), X17
MOV 24(X9), X18
// synthetic carry, one column at a time
MUL X5, X15, X19
MULHU X5, X15, X20
ADD X11, X19 // ADDS X11, X19, X19 (cr=X28)
SLTU X11, X19, X28 // ...
ADD X28, X20 // ADC $0, X20, X20
ADD X6, X19, X15 // ADDS X6, X19, X15 (cr=X28)
SLTU X6, X15, X28 // ...
ADD X28, X20, X6 // ADC $0, X20, X6
MUL X5, X16, X19
MULHU X5, X16, X20
ADD X12, X19 // ADDS X12, X19, X19 (cr=X28)
SLTU X12, X19, X28 // ...
ADD X28, X20 // ADC $0, X20, X20
ADD X6, X19, X16 // ADDS X6, X19, X16 (cr=X28)
SLTU X6, X16, X28 // ...
ADD X28, X20, X6 // ADC $0, X20, X6
MUL X5, X17, X19
MULHU X5, X17, X20
ADD X13, X19 // ADDS X13, X19, X19 (cr=X28)
SLTU X13, X19, X28 // ...
ADD X28, X20 // ADC $0, X20, X20
ADD X6, X19, X17 // ADDS X6, X19, X17 (cr=X28)
SLTU X6, X17, X28 // ...
ADD X28, X20, X6 // ADC $0, X20, X6
MUL X5, X18, X19
MULHU X5, X18, X20
ADD X14, X19 // ADDS X14, X19, X19 (cr=X28)
SLTU X14, X19, X28 // ...
ADD X28, X20 // ADC $0, X20, X20
ADD X6, X19, X18 // ADDS X6, X19, X18 (cr=X28)
SLTU X6, X18, X28 // ...
ADD X28, X20, X6 // ADC $0, X20, X6
MOV X15, 0(X10)
MOV X16, 8(X10)
MOV X17, 16(X10)
MOV X18, 24(X10)
ADD $32, X8
ADD $32, X9
ADD $32, X10
SUB $1, X7
BNEZ X7, loop4cont
loop4done:
MOV X6, c+88(FP)
RET