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
458
|
/* $Id: xform4.S,v 1.1.1.1 2005/07/31 16:46:23 ajax Exp $ */
/*
* Mesa 3-D graphics library
* Version: 3.5
*
* Copyright (C) 1999-2001 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included
* in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
* AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#ifdef USE_X86_64_ASM
#include "matypes.h"
.text
.align 16
.globl _mesa_x86_64_transform_points4_general
_mesa_x86_64_transform_points4_general:
/*
* rdi = dest
* rsi = matrix
* rdx = source
*/
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
.byte 0x66, 0x66, 0x66, 0x90 /* manual align += 3 */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
testl %ecx, %ecx /* verify non-zero count */
prefetchnta 64(%rsi)
jz p4_general_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
prefetch 16(%rdx)
movaps 0(%rsi), %xmm4 /* m3 | m2 | m1 | m0 */
movaps 16(%rsi), %xmm5 /* m7 | m6 | m5 | m4 */
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
movaps 32(%rsi), %xmm6 /* m11 | m10 | m9 | m8 */
movaps 48(%rsi), %xmm7 /* m15 | m14 | m13 | m12 */
p4_general_loop:
movaps (%rdx), %xmm8 /* ox | oy | oz | ow */
prefetchw 16(%rdi)
pshufd $0x00, %xmm8, %xmm0 /* ox | ox | ox | ox */
addq %rax, %rdx
pshufd $0x55, %xmm8, %xmm1 /* oy | oy | oy | oy */
mulps %xmm4, %xmm0 /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
pshufd $0xAA, %xmm8, %xmm2 /* oz | oz | oz | ox */
mulps %xmm5, %xmm1 /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
pshufd $0xFF, %xmm8, %xmm3 /* ow | ow | ow | ow */
mulps %xmm6, %xmm2 /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
addps %xmm1, %xmm0 /* ox*m3+oy*m7 | ... */
mulps %xmm7, %xmm3 /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
addps %xmm2, %xmm0 /* ox*m3+oy*m7+oz*m11 | ... */
prefetch 16(%rdx)
addps %xmm3, %xmm0 /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
movaps %xmm0, (%rdi) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
addq $16, %rdi
decl %ecx
jnz p4_general_loop
p4_general_done:
.byte 0xf3
ret
.section .rodata
.align 16
p4_constants:
.byte 0xff, 0xff, 0xff, 0xff
.byte 0xff, 0xff, 0xff, 0xff
.byte 0xff, 0xff, 0xff, 0xff
.byte 0x00, 0x00, 0x00, 0x00
.byte 0x00, 0x00, 0x00, 0x00
.byte 0x00, 0x00, 0x00, 0x00
.byte 0x00, 0x00, 0x00, 0x00
.float 0f+1.0
.text
.align 16
.globl _mesa_x86_64_transform_points4_3d
/*
* this is slower than _mesa_x86_64_transform_points4_general
* because it ensures that the last matrix row (or is it column?) is 0,0,0,1
*/
_mesa_x86_64_transform_points4_3d:
leaq p4_constants(%rip), %rax
prefetchnta 64(%rsi)
movaps (%rax), %xmm9
movaps 16(%rax), %xmm10
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
testl %ecx, %ecx /* verify non-zero count */
jz p4_3d_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
prefetch 16(%rdx)
movaps 0(%rsi), %xmm4 /* m3 | m2 | m1 | m0 */
movaps 16(%rsi), %xmm5 /* m7 | m6 | m5 | m4 */
andps %xmm9, %xmm4 /* 0.0 | m2 | m1 | m0 */
movaps 32(%rsi), %xmm6 /* m11 | m10 | m9 | m8 */
andps %xmm9, %xmm5 /* 0.0 | m6 | m5 | m4 */
movaps 48(%rsi), %xmm7 /* m15 | m14 | m13 | m12 */
andps %xmm9, %xmm6 /* 0.0 | m10 | m9 | m8 */
andps %xmm9, %xmm7 /* 0.0 | m14 | m13 | m12 */
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
orps %xmm10, %xmm7 /* 1.0 | m14 | m13 | m12 */
p4_3d_loop:
movaps (%rdx), %xmm8 /* ox | oy | oz | ow */
prefetchw 16(%rdi)
pshufd $0x00, %xmm8, %xmm0 /* ox | ox | ox | ox */
addq %rax, %rdx
pshufd $0x55, %xmm8, %xmm1 /* oy | oy | oy | oy */
mulps %xmm4, %xmm0 /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
pshufd $0xAA, %xmm8, %xmm2 /* oz | oz | oz | ox */
mulps %xmm5, %xmm1 /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
pshufd $0xFF, %xmm8, %xmm3 /* ow | ow | ow | ow */
mulps %xmm6, %xmm2 /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
addps %xmm1, %xmm0 /* ox*m3+oy*m7 | ... */
mulps %xmm7, %xmm3 /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
addps %xmm2, %xmm0 /* ox*m3+oy*m7+oz*m11 | ... */
prefetch 16(%rdx)
addps %xmm3, %xmm0 /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
movaps %xmm0, (%rdi) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
addq $16, %rdi
dec %ecx
jnz p4_3d_loop
p4_3d_done:
.byte 0xf3
ret
.align 16
.globl _mesa_x86_64_transform_points4_identity
_mesa_x86_64_transform_points4_identity:
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
test %ecx, %ecx
jz p4_identity_done
movq V4F_START(%rdx), %rsi /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
prefetch 64(%rsi)
prefetchw 64(%rdi)
add %ecx, %ecx
rep movsq
p4_identity_done:
.byte 0xf3
ret
.align 16
.globl _mesa_x86_64_transform_points4_3d_no_rot
_mesa_x86_64_transform_points4_3d_no_rot:
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
test %ecx, %ecx
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
jz p4_3d_no_rot_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
prefetch (%rdx)
movd (%rsi), %mm0 /* | m00 */
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
punpckldq 20(%rsi), %mm0 /* m11 | m00 */
movd 40(%rsi), %mm2 /* | m22 */
movq 48(%rsi), %mm1 /* m31 | m30 */
punpckldq 56(%rsi), %mm2 /* m11 | m00 */
p4_3d_no_rot_loop:
prefetchw 32(%rdi)
movq (%rdx), %mm4 /* x1 | x0 */
movq 8(%rdx), %mm5 /* x3 | x2 */
movd 12(%rdx), %mm7 /* | x3 */
movq %mm5, %mm6 /* x3 | x2 */
pfmul %mm0, %mm4 /* x1*m11 | x0*m00 */
punpckhdq %mm6, %mm6 /* x3 | x3 */
pfmul %mm2, %mm5 /* x3*m32 | x2*m22 */
pfmul %mm1, %mm6 /* x3*m31 | x3*m30 */
pfacc %mm7, %mm5 /* x3 | x2*m22+x3*m32 */
pfadd %mm6, %mm4 /* x1*m11+x3*m31 | x0*m00+x3*m30 */
addq %rax, %rdx
movq %mm4, (%rdi) /* write r0, r1 */
movq %mm5, 8(%rdi) /* write r2, r3 */
addq $16, %rdi
decl %ecx
prefetch 32(%rdx)
jnz p4_3d_no_rot_loop
p4_3d_no_rot_done:
femms
ret
.align 16
.globl _mesa_x86_64_transform_points4_perspective
_mesa_x86_64_transform_points4_perspective:
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
test %ecx, %ecx
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
jz p4_perspective_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
movd (%rsi), %mm0 /* | m00 */
pxor %mm7, %mm7 /* 0 | 0 */
punpckldq 20(%rsi), %mm0 /* m11 | m00 */
movq 32(%rsi), %mm2 /* m21 | m20 */
prefetch (%rdx)
movd 40(%rsi), %mm1 /* | m22 */
.byte 0x66, 0x66, 0x90 /* manual align += 3 */
punpckldq 56(%rsi), %mm1 /* m32 | m22 */
p4_perspective_loop:
prefetchw 32(%rdi) /* prefetch 2 vertices ahead */
movq (%rdx), %mm4 /* x1 | x0 */
movq 8(%rdx), %mm5 /* x3 | x2 */
movd 8(%rdx), %mm3 /* | x2 */
movq %mm5, %mm6 /* x3 | x2 */
pfmul %mm0, %mm4 /* x1*m11 | x0*m00 */
punpckldq %mm5, %mm5 /* x2 | x2 */
pfmul %mm2, %mm5 /* x2*m21 | x2*m20 */
pfsubr %mm7, %mm3 /* | -x2 */
pfmul %mm1, %mm6 /* x3*m32 | x2*m22 */
pfadd %mm4, %mm5 /* x1*m11+x2*m21 | x0*m00+x2*m20 */
pfacc %mm3, %mm6 /* -x2 | x2*m22+x3*m32 */
movq %mm5, (%rdi) /* write r0, r1 */
addq %rax, %rdx
movq %mm6, 8(%rdi) /* write r2, r3 */
addq $16, %rdi
decl %ecx
prefetch 32(%rdx) /* hopefully stride is zero */
jnz p4_perspective_loop
p4_perspective_done:
femms
ret
.align 16
.globl _mesa_x86_64_transform_points4_2d_no_rot
_mesa_x86_64_transform_points4_2d_no_rot:
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
test %ecx, %ecx
.byte 0x90 /* manual align += 1 */
jz p4_2d_no_rot_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
movd (%rsi), %mm0 /* | m00 */
prefetch (%rdx)
punpckldq 20(%rsi), %mm0 /* m11 | m00 */
movq 48(%rsi), %mm1 /* m31 | m30 */
p4_2d_no_rot_loop:
prefetchw 32(%rdi) /* prefetch 2 vertices ahead */
movq (%rdx), %mm4 /* x1 | x0 */
movq 8(%rdx), %mm5 /* x3 | x2 */
pfmul %mm0, %mm4 /* x1*m11 | x0*m00 */
movq %mm5, %mm6 /* x3 | x2 */
punpckhdq %mm6, %mm6 /* x3 | x3 */
addq %rax, %rdx
pfmul %mm1, %mm6 /* x3*m31 | x3*m30 */
prefetch 32(%rdx) /* hopefully stride is zero */
pfadd %mm4, %mm6 /* x1*m11+x3*m31 | x0*m00+x3*m30 */
movq %mm6, (%rdi) /* write r0, r1 */
movq %mm5, 8(%rdi) /* write r2, r3 */
addq $16, %rdi
decl %ecx
jnz p4_2d_no_rot_loop
p4_2d_no_rot_done:
femms
ret
.align 16
.globl _mesa_x86_64_transform_points4_2d
_mesa_x86_64_transform_points4_2d:
movl V4F_COUNT(%rdx), %ecx /* count */
movzx V4F_STRIDE(%rdx), %eax /* stride */
movl %ecx, V4F_COUNT(%rdi) /* set dest count */
movl $4, V4F_SIZE(%rdi) /* set dest size */
.byte 0x66, 0x66, 0x90 /* manual align += 4 */
orl $VEC_SIZE_4, V4F_FLAGS(%rdi)/* set dest flags */
test %ecx, %ecx
.byte 0x66, 0x66, 0x90 /* manual align += 4 */
jz p4_2d_done
movq V4F_START(%rdx), %rdx /* ptr to first src vertex */
movq V4F_START(%rdi), %rdi /* ptr to first dest vertex */
movd (%rsi), %mm0 /* | m00 */
movd 4(%rsi), %mm1 /* | m01 */
prefetch (%rdx)
punpckldq 16(%rsi), %mm0 /* m10 | m00 */
.byte 0x66, 0x66, 0x90 /* manual align += 4 */
punpckldq 20(%rsi), %mm1 /* m11 | m01 */
movq 48(%rsi), %mm2 /* m31 | m30 */
p4_2d_loop:
prefetchw 32(%rdi) /* prefetch 2 vertices ahead */
movq (%rdx), %mm3 /* x1 | x0 */
movq 8(%rdx), %mm5 /* x3 | x2 */
movq %mm3, %mm4 /* x1 | x0 */
movq %mm5, %mm6 /* x3 | x2 */
pfmul %mm1, %mm4 /* x1*m11 | x0*m01 */
punpckhdq %mm6, %mm6 /* x3 | x3 */
pfmul %mm0, %mm3 /* x1*m10 | x0*m00 */
addq %rax, %rdx
pfacc %mm4, %mm3 /* x0*m01+x1*m11 | x0*m00+x1*m10 */
pfmul %mm2, %mm6 /* x3*m31 | x3*m30 */
prefetch 32(%rdx) /* hopefully stride is zero */
pfadd %mm6, %mm3 /* r1 | r0 */
movq %mm3, (%rdi) /* write r0, r1 */
movq %mm5, 8(%rdi) /* write r2, r3 */
addq $16, %rdi
decl %ecx
jnz p4_2d_loop
p4_2d_done:
femms
ret
#endif
|
