Debian Codegraph @

src/shaders/post_processing/gen8/Save_AVS_PL3.g8a

raw
1 /*
2 * Copyright 2000-2011 Intel Corporation All Rights Reserved
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 * Author: Zhao Yakui <[email protected]>
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17 // 84 // Total instruction count
18 // 1 // Total kernel count
19
20
21
22 // Module name: common.inc
23 //
24 // Common header file for all Video-Processing kernels
25 //
26
27 .default_execution_size (16)
28 .default_register_type :ub
29
30 .reg_count_total 128
31 .reg_count_payload 7
32
33 //========== Common constants ==========
34
35
36 //========== Macros ==========
37
38
39 //Fast Jump, For more details see "Set_Layer_N.asm"
40
41
42 //========== Defines ====================
43
44 //========== Static Parameters (Common To All) ==========
45 //r1
46
47
48 //r2
49
50 // e.g. byte0 byte1 byte2
51 // YUYV 0 1 3
52 // YVYU 0 3 1
53
54 //Color Pipe (IECP) parameters
55
56
57 //ByteCopy
58
59
60 //r4
61
62 // e.g. byte0 byte1 byte2
63 // YUYV 0 1 3
64 // YVYU 0 3 1
65
66
67 //========== Inline parameters (Common To All) ===========
68
69
70 //============== Binding Index Table===========
71 //Common between DNDI and DNUV
72
73
74 //================= Common Message Descriptor =====
75 // Message descriptor for thread spawning
76 // Message Descriptors
77 // = 000 0001 (min message len 1 ) 0,0000 (resp len 0 -add later)
78 // 0000,0000,0000
79 // 0001(Spawn a root thread),0001 (Root thread spawn thread)
80 // = 0x02000011
81 // Thread Spawner Message Descriptor
82
83
84 // Message descriptor for atomic operation add
85 // Message Descriptors
86 // = 000 0110 (min message len 6 ) 0,0000 (resp len 0 -add later)
87 // 1(header present)001,10(typed atomic operation)0(return enabled)0(slot group, low 8 bits),0111 (AOP_Add)
88 // 0000,0000 (Binding table index, added later)
89 // = 0x02000011
90
91 // Atomic Operation Add Message Descriptor
92
93
94 // Message descriptor for dataport media write
95 // Message Descriptors
96 // = 000 0001 (min message len 1 - add later) 00000 (resp len 0)
97 // 1 (header present 1) 0 1010 (media block write) 000000
98 // 00000000 (binding table index - set later)
99 // = 0x020A8000
100
101
102 // Message Length defines
103
104
105 // Response Length defines
106
107
108 // Block Width and Height Size defines
109
110
111 // Extended Message Descriptors
112
113
114 // Common message descriptors:
115
116
117 //===================== Math Function Control ===================================
118
119
120 //============ Message Registers ===============
121 // buf4 starts from r28
122
123
124 //#define mMSGHDR_EOT r43 // Dummy Message Register for EOT
125
126
127 .declare mubMSGPAYLOAD Base=r30 ElementSize=1 SrcRegion=<16;16,1> Type=ub
128 .declare muwMSGPAYLOAD Base=r30 ElementSize=2 SrcRegion=<16;16,1> Type=uw
129 .declare mudMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=ud
130 .declare mfMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=f
131
132 //=================== End of thread instruction ===========================
133
134
135 //=====================Pointers Used=====================================
136
137
138 //=======================================================================
139
140
141 //r11-r17
142 // Define temp space for any usages
143
144
145 // Common Buffers
146
147
148 // temp space for rotation
149
150 .declare fROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
151
152 .declare udROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
153
154 .declare uwROBUF Base=r11.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
155
156 .declare ubROBUF Base=r11.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
157
158 .declare ub4ROBUF Base=r11.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
159
160
161 // End of common.inc
162
163
164 // Module name: Save_AVS_PL3.asm
165 //
166 // Save PL3 420 frame data block of size 16x16
167 //
168 // To save 16x16 block (16x16 byte of Y and 8x8 byte of U and V each) we need 3 send instructions with one of size 16x16 and two of size 8x8.
169 // -----------------
170 // | 16x16 Y |
171 // | |
172 // -----------------
173 // | 8x8 U |
174 // ---------
175 // | 8x8 V |
176 // ---------
177
178 //-----------------------------------------------------------------
179 //The layout of data is as follows:
180 //mMSGHDR0 : Y data header (16x16)
181 //mubMSGPAYLOAD0 : Y data payload (8 GRFs)
182 //mMSGHDR1 : U data header (8x8)
183 //mubMSGPAYLOAD1 : U data payload (2 GRFs)
184 //mMSGHDR2 : V data header (8x8)
185 //mubMSGPAYLOAD2 : V data payload (2 GRFs)
186 //------------------------------------------------------------------
187
188
189
190 // Module name: Save.inc
191
192
193
194
195 // Description: Includes all definitions explicit to Fast Composite.
196
197
198
199
200 // End of common.inc
201
202
203 //========== GRF partition ==========
204 // r0 header : r0 (1 GRF)
205 // Static parameters : r1 - r6 (6 GRFS)
206 // Inline parameters : r7 - r8 (2 GRFs)
207 // MSGSRC : r27 (1 GRF)
208 //===================================
209
210 //Interface:
211 //========== Static Parameters (Explicit To Fast Composite) ==========
212 //r1
213 //CSC Set 0
214
215
216 .declare udCSC_CURBE Base=r1.0 ElementSize=4 Type=ud
217
218 //Constant alpha
219
220
221 //r2
222
223
224 // Gen7 AVS WA
225
226
227 // WiDi Definitions
228
229
230 //Colorfill
231
232
233 // 0: 0-degree, 1: 90, 2: 180, 3: 270-degree, clockwise.
234
235 .declare ubCOLOR_PIXEL_VAL Base=r2.20 ElementSize=1 SrcRegion=<0;1,0> DstRegion=<1> Type=ub
236
237 //r3
238 //Normalised Ratio of Horizontal step size with main video for all layers
239
240
241 //Normalised Ratio of Horizontal step size with main video for all layers becomes
242 //Normalised Horizontal step size for all layers in VP_Setup.asm
243
244
245 //r4
246 //Normalised Vertical step size for all layers
247
248
249 //r5
250 //Normalised Vertical Frame Origin for all layers
251
252
253 //r6
254 //Normalised Horizontal Frame Origin for all layers
255
256
257 //========== Inline Parameters (Explicit To Fast Composite) ==========
258
259
260 //Main video Step X
261
262
263 //====================== Binding table (Explicit To Fast Composite)=========================================
264
265
266 //Used by Interlaced Scaling Kernels
267
268
269 //========== Sampler State Table Index (Explicit To Fast Composite)==========
270 //Sampler Index for AVS/IEF messages
271
272
273 //Sampler Index for SIMD16 sampler messages
274
275
276 //=============================================================================
277
278 .declare fBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
279 .declare fBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
280 .declare fBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
281 .declare fBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
282 .declare fBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
283 .declare fBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
284
285 .declare udBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
286 .declare udBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
287 .declare udBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
288 .declare udBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
289 .declare udBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
290 .declare udBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
291
292 .declare uwBUFFER_0 Base=r64.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
293 .declare uwBUFFER_1 Base=r80.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
294 .declare uwBUFFER_2 Base=r96.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
295 .declare uwBUFFER_3 Base=r112.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
296 .declare uwBUFFER_4 Base=r28.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
297 .declare uwBUFFER_5 Base=r46.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
298
299 .declare ubBUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
300 .declare ubBUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
301 .declare ubBUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
302 .declare ubBUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
303 .declare ubBUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
304 .declare ubBUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
305
306 .declare ub4BUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
307 .declare ub4BUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
308 .declare ub4BUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
309 .declare ub4BUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
310 .declare ub4BUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
311 .declare ub4BUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
312
313 //Pointer to mask reg
314
315
316 //r18
317
318
319 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
320 // NODDCLR, NODDCHK flags. -rT
321
322
323 .declare udCSC_COEFF_0 Base=r18.0 ElementSize=4 Type=ud // 1 GRF
324
325 //r19
326
327
328 .declare udCSC_COEFF_1 Base=r19.0 ElementSize=4 Type=ud // 1 GRF
329
330
331 //r20
332
333 .declare uwALPHA_MASK_REG_TEMP Base=r20.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
334
335 //r21
336
337 .declare uwALPHA_MASK_REG Base=r21.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
338
339 //r22
340
341
342 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
343 // NODDCLR, NODDCHK flags. -rT
344
345
346 //Keep fORIGIN_X_NLAS, fY_OFFSET_2ND_BLOCK, fSTEP_X_NLAS, pMSGDSC_COPY, ubCONST_ALPHA_COPY as
347 //sub registers of same GRF to enable using NODDCLR NODDCHK. -rT
348
349 //r23
350
351
352 //Lumakey
353
354
355 //r24
356
357
358 //r25
359
360
361 //r26
362
363
364 //defines to generate LABELS during compile time.
365
366
367 //Msg payload buffers; upto 4 full-size messages can be written
368
369
370 .declare mudMSGPAYLOAD0 Base=r29.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
371 .declare mudMSGPAYLOAD1 Base=r38.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
372 .declare mudMSGPAYLOAD2 Base=r47.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
373 .declare mudMSGPAYLOAD3 Base=r56.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
374
375 .declare muwMSGPAYLOAD0 Base=r29.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
376 .declare muwMSGPAYLOAD1 Base=r38.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
377 .declare muwMSGPAYLOAD2 Base=r47.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
378 .declare muwMSGPAYLOAD3 Base=r56.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
379
380 .declare mubMSGPAYLOAD0 Base=r29.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
381 .declare mubMSGPAYLOAD1 Base=r38.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
382 .declare mubMSGPAYLOAD2 Base=r47.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
383 .declare mubMSGPAYLOAD3 Base=r56.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
384 .declare mubMSGPAYLOAD4 Base=r32.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
385 .declare mubMSGPAYLOAD5 Base=r41.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
386 .declare mubMSGPAYLOAD6 Base=r50.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
387 .declare mubMSGPAYLOAD7 Base=r59.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
388
389
390 // the r17 register (nTEMP0) is originally defined from "Common.inc"
391 // instead of re-defining a nTEMP0 here, we use "SAVE_RGB" suffix for its naming
392
393 .declare uwTemp0 Base=r17.0 ElementSize=2 Type=uw
394
395
396 //_SAVE_INC_
397
398
399 // At the save module we have all 8 address sub-registers available.
400 // So we will use PING-PONG type of scheme to save the data using
401 // pointers pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4. This will help
402 // reduce dependency. - rT
403
404 //wBUFF_CHNL_PTR points to either buffer 0 or buffer 4.
405 //Add appropriate offsets to get pointers for all buffers (1,2,3 or 5).
406 //Offsets are zero for buffer 0 and buffer 4.
407 add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
408 add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
409
410 //Set up header for Y,U and V data
411 mov (8) r28<1>:ud r27<8;8,1>:ud
412 mov (8) r37<1>:ud r27<8;8,1>:ud
413 mov (8) r46<1>:ud r27<8;8,1>:ud
414
415 mov (2) r28.0<1>:d r9.0<2;2,1>:w { NoDDClr } //ORI Y (LUMA) = ORI
416 shr (2) r37.0<1>:d r9.0<2;2,1>:w 1:w { NoDDClr } //H/V ORI U = H/V ORI/2
417 shr (2) r46.0<1>:d r9.0<2;2,1>:w 1:w { NoDDClr } //H/V ORI V = H/V ORI/2
418
419 mov (1) r28.2<1>:ud 0xF000F:ud { NoDDChk } // Y Block width and height (16x16)
420 mov (1) r37.2<1>:ud 0x70007:ud { NoDDChk } // U Block width and height (8x8)
421 mov (1) r46.2<1>:ud 0x70007:ud { NoDDChk } // V Block width and height (8x8)
422
423 // Unscramble, and pack data directly to MRFs
424
425 // Data 16x16 block is divided as -
426 // ---------
427 // | 0 |
428 // ---------
429 // | 1 |
430 // ---------
431 // | 2 |
432 // ---------
433 // | 3 |
434 // ---------
435 // All sub-blocks are of size 16x4
436 // 0: ubBUFFER_0
437 // 1: ubBUFFER_1, ubBUFFER_0+16
438 // 2: ubBUFFER_2
439 // 3: ubBUFFER_3, ubBUFFER_2+16
440
441 //Y Rounding, first
442 add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
443 add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
444 add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
445 add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
446
447 // U rounding
448 add.sat (8) r[a0.2,0]<2>:uw r[a0.2,0]<16;8,2>:uw 0x0080:uw
449 add.sat (8) r[a0.2,64]<2>:uw r[a0.2,64]<16;8,2>:uw 0x0080:uw
450
451 // V rounding
452 add.sat (8) r[a0.0,0]<2>:uw r[a0.0,0]<16;8,2>:uw 0x0080:uw
453 add.sat (8) r[a0.0,64]<2>:uw r[a0.0,64]<16;8,2>:uw 0x0080:uw
454
455 add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
456
457 //Y Rounding, second
458 add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
459 add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
460 add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
461 add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
462
463 // U rounding
464 add.sat (8) r[a0.6,0]<2>:uw r[a0.6,0]<16;8,2>:uw 0x0080:uw
465 add.sat (8) r[a0.6,64]<2>:uw r[a0.6,64]<16;8,2>:uw 0x0080:uw
466
467 // V rounding
468 add.sat (8) r[a0.4,0]<2>:uw r[a0.4,0]<16;8,2>:uw 0x0080:uw
469 add.sat (8) r[a0.4,64]<2>:uw r[a0.4,64]<16;8,2>:uw 0x0080:uw
470
471 add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
472
473 //Y Rounding, third
474 add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
475 add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
476 add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
477 add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
478
479 // U rounding
480 add.sat (8) r[a0.2,0]<2>:uw r[a0.2,0]<16;8,2>:uw 0x0080:uw
481 add.sat (8) r[a0.2,64]<2>:uw r[a0.2,64]<16;8,2>:uw 0x0080:uw
482
483 // V rounding
484 add.sat (8) r[a0.0,0]<2>:uw r[a0.0,0]<16;8,2>:uw 0x0080:uw
485 add.sat (8) r[a0.0,64]<2>:uw r[a0.0,64]<16;8,2>:uw 0x0080:uw
486
487
488 //Y Rounding, fourth
489 add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
490 add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
491 add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
492 add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
493
494 // U rounding
495 add.sat (8) r[a0.6,0]<2>:uw r[a0.6,0]<16;8,2>:uw 0x0080:uw
496 add.sat (8) r[a0.6,64]<2>:uw r[a0.6,64]<16;8,2>:uw 0x0080:uw
497
498 // V rounding
499 add.sat (8) r[a0.4,0]<2>:uw r[a0.4,0]<16;8,2>:uw 0x0080:uw
500 add.sat (8) r[a0.4,64]<2>:uw r[a0.4,64]<16;8,2>:uw 0x0080:uw
501
502 // restore the TOP and BOT pointers
503 add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
504 add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
505
506 //Buffer 0
507 //Move Y to msg payload
508 mov (16) mubMSGPAYLOAD0(0,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
509 mov (16) mubMSGPAYLOAD0(0,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
510 mov (16) mubMSGPAYLOAD0(1,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
511 mov (16) mubMSGPAYLOAD0(1,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
512
513 //Move U to msg payload
514 mov (8) mubMSGPAYLOAD1(0,0)<1> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
515 mov (8) mubMSGPAYLOAD1(0,8)<1> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
516
517 //Move V to msg payload
518 mov (8) mubMSGPAYLOAD2(0,0)<1> r[a0.0, 1]<32;8,4>:ub { NoDDClr }
519 mov (8) mubMSGPAYLOAD2(0,8)<1> r[a0.0, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
520
521 add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
522
523 //Buffer 1
524 mov (16) mubMSGPAYLOAD0(2,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
525 mov (16) mubMSGPAYLOAD0(2,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
526 mov (16) mubMSGPAYLOAD0(3,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
527 mov (16) mubMSGPAYLOAD0(3,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
528
529 mov (8) mubMSGPAYLOAD1(0,16)<1> r[a0.6, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
530 mov (8) mubMSGPAYLOAD1(0,24)<1> r[a0.6, 65]<32;8,4>:ub { NoDDChk }
531
532 mov (8) mubMSGPAYLOAD2(0,16)<1> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
533 mov (8) mubMSGPAYLOAD2(0,24)<1> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
534
535 add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
536
537 //Buffer 2
538 mov (16) mubMSGPAYLOAD0(4,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
539 mov (16) mubMSGPAYLOAD0(4,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
540 mov (16) mubMSGPAYLOAD0(5,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
541 mov (16) mubMSGPAYLOAD0(5,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
542
543 mov (8) mubMSGPAYLOAD1(1,0)<1> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
544 mov (8) mubMSGPAYLOAD1(1,8)<1> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
545
546 mov (8) mubMSGPAYLOAD2(1,0)<1> r[a0.0, 1]<32;8,4>:ub { NoDDClr }
547 mov (8) mubMSGPAYLOAD2(1,8)<1> r[a0.0, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
548
549 //Buffer 3
550 mov (16) mubMSGPAYLOAD0(6,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
551 mov (16) mubMSGPAYLOAD0(6,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
552 mov (16) mubMSGPAYLOAD0(7,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
553 mov (16) mubMSGPAYLOAD0(7,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
554
555 mov (8) mubMSGPAYLOAD1(1,16)<1> r[a0.6, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
556 mov (8) mubMSGPAYLOAD1(1,24)<1> r[a0.6, 65]<32;8,4>:ub { NoDDChk }
557
558 mov (8) mubMSGPAYLOAD2(1,16)<1> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
559 mov (8) mubMSGPAYLOAD2(1,24)<1> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
560
561 //===========================================================================
562
563 send (1) null<1>:d r28 0xc 0x120A8018:ud
564 send (1) null<1>:d r37 0xc 0x60A8019:ud
565 send (1) null<1>:d r46 0xc 0x60A801A:ud