Debian Codegraph @

src/shaders/post_processing/gen8/Save_AVS_NV12.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 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16 // 131 // Total instruction count
17 // 1 // Total kernel count
18
19
20
21 // Module name: common.inc
22 //
23 // Common header file for all Video-Processing kernels
24 //
25
26 .default_execution_size (16)
27 .default_register_type :ub
28
29 .reg_count_total 128
30 .reg_count_payload 7
31
32 //========== Common constants ==========
33
34
35 //========== Macros ==========
36
37
38 //Fast Jump, For more details see "Set_Layer_N.asm"
39
40
41 //========== Defines ====================
42
43 //========== Static Parameters (Common To All) ==========
44 //r1
45
46
47 //r2
48
49 // e.g. byte0 byte1 byte2
50 // YUYV 0 1 3
51 // YVYU 0 3 1
52
53 //Color Pipe (IECP) parameters
54
55
56 //ByteCopy
57
58
59 //r4
60
61 // e.g. byte0 byte1 byte2
62 // YUYV 0 1 3
63 // YVYU 0 3 1
64
65
66 //========== Inline parameters (Common To All) ===========
67
68
69 //============== Binding Index Table===========
70 //Common between DNDI and DNUV
71
72
73 //================= Common Message Descriptor =====
74 // Message descriptor for thread spawning
75 // Message Descriptors
76 // = 000 0001 (min message len 1 ) 0,0000 (resp len 0 -add later)
77 // 0000,0000,0000
78 // 0001(Spawn a root thread),0001 (Root thread spawn thread)
79 // = 0x02000011
80 // Thread Spawner Message Descriptor
81
82
83 // Message descriptor for atomic operation add
84 // Message Descriptors
85 // = 000 0110 (min message len 6 ) 0,0000 (resp len 0 -add later)
86 // 1(header present)001,10(typed atomic operation)0(return enabled)0(slot group, low 8 bits),0111 (AOP_Add)
87 // 0000,0000 (Binding table index, added later)
88 // = 0x02000011
89
90 // Atomic Operation Add Message Descriptor
91
92
93 // Message descriptor for dataport media write
94 // Message Descriptors
95 // = 000 0001 (min message len 1 - add later) 00000 (resp len 0)
96 // 1 (header present 1) 0 1010 (media block write) 000000
97 // 00000000 (binding table index - set later)
98 // = 0x020A8000
99
100
101 // Message Length defines
102
103
104 // Response Length defines
105
106
107 // Block Width and Height Size defines
108
109
110 // Extended Message Descriptors
111
112
113 // Common message descriptors:
114
115
116 //===================== Math Function Control ===================================
117
118
119 //============ Message Registers ===============
120 // buf4 starts from r28
121
122
123 //#define mMSGHDR_EOT r43 // Dummy Message Register for EOT
124
125
126 .declare mubMSGPAYLOAD Base=r30 ElementSize=1 SrcRegion=<16;16,1> Type=ub
127 .declare muwMSGPAYLOAD Base=r30 ElementSize=2 SrcRegion=<16;16,1> Type=uw
128 .declare mudMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=ud
129 .declare mfMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=f
130
131 //=================== End of thread instruction ===========================
132
133
134 //=====================Pointers Used=====================================
135
136
137 //=======================================================================
138
139
140 //r11-r17
141 // Define temp space for any usages
142
143
144 // Common Buffers
145
146
147 // temp space for rotation
148
149 .declare fROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
150
151 .declare udROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
152
153 .declare uwROBUF Base=r11.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
154
155 .declare ubROBUF Base=r11.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
156
157 .declare ub4ROBUF Base=r11.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
158
159
160 // End of common.inc
161
162
163 // Module name: Save_AVS_NV12.asm
164 //
165 // Save NV12 420 frame data block of size 16x16
166 //
167 // To save 16x16 block (16x16 bytes of Y and 16x8 bytes of interleaved UV), we need 2 send instructions with of size 16x16 and 16x8 each.
168 // ---------------
169 // | 16x16 |
170 // | YUYV |
171 // ---------------
172 // | 16x8 UV |
173 // ---------------
174
175 //-----------------------------------------------------------------
176 //The layout of data is as follows:
177 //mMSGHDR0 : Y data header (16x16)
178 //mubMSGPAYLOAD0 : Y data payload (8 GRFs)
179 //mMSGHDR1 : U data header (16x8)
180 //mubMSGPAYLOAD1 : U data payload (4 GRFs)
181 //------------------------------------------------------------------
182
183
184
185 // Module name: Save.inc
186
187
188
189
190 // Description: Includes all definitions explicit to Fast Composite.
191
192
193
194
195 // End of common.inc
196
197
198 //========== GRF partition ==========
199 // r0 header : r0 (1 GRF)
200 // Static parameters : r1 - r6 (6 GRFS)
201 // Inline parameters : r7 - r8 (2 GRFs)
202 // MSGSRC : r27 (1 GRF)
203 //===================================
204
205 //Interface:
206 //========== Static Parameters (Explicit To Fast Composite) ==========
207 //r1
208 //CSC Set 0
209
210
211 .declare udCSC_CURBE Base=r1.0 ElementSize=4 Type=ud
212
213 //Constant alpha
214
215
216 //r2
217
218
219 // Gen7 AVS WA
220
221
222 // WiDi Definitions
223
224
225 //Colorfill
226
227
228 // 0: 0-degree, 1: 90, 2: 180, 3: 270-degree, clockwise.
229
230 .declare ubCOLOR_PIXEL_VAL Base=r2.20 ElementSize=1 SrcRegion=<0;1,0> DstRegion=<1> Type=ub
231
232 //r3
233 //Normalised Ratio of Horizontal step size with main video for all layers
234
235
236 //Normalised Ratio of Horizontal step size with main video for all layers becomes
237 //Normalised Horizontal step size for all layers in VP_Setup.asm
238
239
240 //r4
241 //Normalised Vertical step size for all layers
242
243
244 //r5
245 //Normalised Vertical Frame Origin for all layers
246
247
248 //r6
249 //Normalised Horizontal Frame Origin for all layers
250
251
252 //========== Inline Parameters (Explicit To Fast Composite) ==========
253
254
255 //Main video Step X
256
257
258 //====================== Binding table (Explicit To Fast Composite)=========================================
259
260
261 //Used by Interlaced Scaling Kernels
262
263
264 //========== Sampler State Table Index (Explicit To Fast Composite)==========
265 //Sampler Index for AVS/IEF messages
266
267
268 //Sampler Index for SIMD16 sampler messages
269
270
271 //=============================================================================
272
273 .declare fBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
274 .declare fBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
275 .declare fBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
276 .declare fBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
277 .declare fBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
278 .declare fBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
279
280 .declare udBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
281 .declare udBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
282 .declare udBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
283 .declare udBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
284 .declare udBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
285 .declare udBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
286
287 .declare uwBUFFER_0 Base=r64.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
288 .declare uwBUFFER_1 Base=r80.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
289 .declare uwBUFFER_2 Base=r96.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
290 .declare uwBUFFER_3 Base=r112.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
291 .declare uwBUFFER_4 Base=r28.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
292 .declare uwBUFFER_5 Base=r46.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
293
294 .declare ubBUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
295 .declare ubBUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
296 .declare ubBUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
297 .declare ubBUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
298 .declare ubBUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
299 .declare ubBUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
300
301 .declare ub4BUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
302 .declare ub4BUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
303 .declare ub4BUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
304 .declare ub4BUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
305 .declare ub4BUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
306 .declare ub4BUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
307
308 //Pointer to mask reg
309
310
311 //r18
312
313
314 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
315 // NODDCLR, NODDCHK flags. -rT
316
317
318 .declare udCSC_COEFF_0 Base=r18.0 ElementSize=4 Type=ud // 1 GRF
319
320 //r19
321
322
323 .declare udCSC_COEFF_1 Base=r19.0 ElementSize=4 Type=ud // 1 GRF
324
325
326 //r20
327
328 .declare uwALPHA_MASK_REG_TEMP Base=r20.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
329
330 //r21
331
332 .declare uwALPHA_MASK_REG Base=r21.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
333
334 //r22
335
336
337 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
338 // NODDCLR, NODDCHK flags. -rT
339
340
341 //Keep fORIGIN_X_NLAS, fY_OFFSET_2ND_BLOCK, fSTEP_X_NLAS, pMSGDSC_COPY, ubCONST_ALPHA_COPY as
342 //sub registers of same GRF to enable using NODDCLR NODDCHK. -rT
343
344 //r23
345
346
347 //Lumakey
348
349
350 //r24
351
352
353 //r25
354
355
356 //r26
357
358
359 //defines to generate LABELS during compile time.
360
361
362 //Msg payload buffers; upto 4 full-size messages can be written
363
364
365 .declare mudMSGPAYLOAD0 Base=r29.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
366 .declare mudMSGPAYLOAD1 Base=r38.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
367 .declare mudMSGPAYLOAD2 Base=r47.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
368 .declare mudMSGPAYLOAD3 Base=r56.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
369
370 .declare muwMSGPAYLOAD0 Base=r29.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
371 .declare muwMSGPAYLOAD1 Base=r38.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
372 .declare muwMSGPAYLOAD2 Base=r47.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
373 .declare muwMSGPAYLOAD3 Base=r56.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
374
375 .declare mubMSGPAYLOAD0 Base=r29.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
376 .declare mubMSGPAYLOAD1 Base=r38.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
377 .declare mubMSGPAYLOAD2 Base=r47.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
378 .declare mubMSGPAYLOAD3 Base=r56.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
379 .declare mubMSGPAYLOAD4 Base=r32.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
380 .declare mubMSGPAYLOAD5 Base=r41.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
381 .declare mubMSGPAYLOAD6 Base=r50.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
382 .declare mubMSGPAYLOAD7 Base=r59.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
383
384
385 // the r17 register (nTEMP0) is originally defined from "Common.inc"
386 // instead of re-defining a nTEMP0 here, we use "SAVE_RGB" suffix for its naming
387
388 .declare uwTemp0 Base=r17.0 ElementSize=2 Type=uw
389
390
391 //_SAVE_INC_
392
393
394 // At the save module we have all 8 address sub-registers available.
395 // So we will use PING-PONG type of scheme to save the data using
396 // pointers pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4. This will help
397 // reduce dependency. - rT
398
399 //wBUFF_CHNL_PTR points to either buffer 0 or buffer 4.
400 //Add appropriate offsets to get pointers for all buffers (1,2,3 or 5).
401 //Offsets are zero for buffer 0 and buffer 4.
402 add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
403 add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
404
405 //Set up header for Y,U and V data
406 mov (8) r28<1>:ud r27<8;8,1>:ud
407 mov (8) r37<1>:ud r27<8;8,1>:ud
408
409 mov (2) r28.0<1>:d r9.0<2;2,1>:w { NoDDClr } //ORI Y (LUMA) = ORI
410 mov (1) r37.0<1>:d r9.0<0;1,0>:w { NoDDClr } //H ORI (CHROMA) = H ORI
411 shr (1) r37.1<1>:d r9.1<0;1,0>:w 1:w { NoDDClr, NoDDChk } //V ORI (CHROMA) = V ORI/2
412
413 mov (1) r28.2<1>:ud 0xF000F:ud { NoDDChk } // Y Block width and height (16x16)
414 mov (1) r37.2<1>:ud 0x7000F:ud { NoDDChk } // UV Block width and height(16x8)
415
416 // Unscramble, and pack data directly to MRFs
417
418 // Data 16x16 block is divided as -
419 // ---------
420 // | 0 |
421 // ---------
422 // | 1 |
423 // ---------
424 // | 2 |
425 // ---------
426 // | 3 |
427 // ---------
428 // All sub-blocks are of size 16x4
429 // 0: ubBUFFER_0
430 // 1: ubBUFFER_1, ubBUFFER_0+16
431 // 2: ubBUFFER_2
432 // 3: ubBUFFER_3, ubBUFFER_2+16
433
434 //Y Rounding 16x4 top part
435 add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
436 add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
437 add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
438 add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
439
440 // U Averaging and Rounding, 8x2 top part
441 add.sat (8) r[a0.2,0]<2>:uw r[a0.2,0]<16;8,2>:uw 0x0080:uw
442 add.sat (8) r[a0.2,64]<2>:uw r[a0.2,64]<16;8,2>:uw 0x0080:uw
443
444 // V Averaging and Rounding, 8x2 top part
445 add.sat (8) r[a0.0,0]<2>:uw r[a0.0,0]<16;8,2>:uw 0x0080:uw
446 add.sat (8) r[a0.0,64]<2>:uw r[a0.0,64]<16;8,2>:uw 0x0080:uw
447
448 add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
449
450 //Y Rounding, 16x4 bottom part
451 add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
452 add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
453 add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
454 add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
455
456 // U Averaging and Rounding, 8x2 bottom part
457 add.sat (8) r[a0.6,0]<2>:uw r[a0.6,0]<16;8,2>:uw 0x0080:uw
458 add.sat (8) r[a0.6,64]<2>:uw r[a0.6,64]<16;8,2>:uw 0x0080:uw
459
460 // V Averaging and Rounding, 8x2 bottom part
461 add.sat (8) r[a0.4,0]<2>:uw r[a0.4,0]<16;8,2>:uw 0x0080:uw
462 add.sat (8) r[a0.4,64]<2>:uw r[a0.4,64]<16;8,2>:uw 0x0080:uw
463
464 add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
465
466 //Y Rounding 16x4 top part
467 add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
468 add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
469 add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
470 add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
471
472 // U Averaging and Rounding, 8x2 top part
473 add.sat (8) r[a0.2,0]<2>:uw r[a0.2,0]<16;8,2>:uw 0x0080:uw
474 add.sat (8) r[a0.2,64]<2>:uw r[a0.2,64]<16;8,2>:uw 0x0080:uw
475
476 // V Averaging and Rounding, 8x2 top part
477 add.sat (8) r[a0.0,0]<2>:uw r[a0.0,0]<16;8,2>:uw 0x0080:uw
478 add.sat (8) r[a0.0,64]<2>:uw r[a0.0,64]<16;8,2>:uw 0x0080:uw
479
480 //Y Rounding, 16x4 bottom part
481 add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
482 add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
483 add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
484 add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
485
486 // U Averaging and Rounding, 8x2 bottom part
487 add.sat (8) r[a0.6,0]<2>:uw r[a0.6,0]<16;8,2>:uw 0x0080:uw
488 add.sat (8) r[a0.6,64]<2>:uw r[a0.6,64]<16;8,2>:uw 0x0080:uw
489
490 // V Averaging and Rounding, 8x2 bottom part
491 add.sat (8) r[a0.4,0]<2>:uw r[a0.4,0]<16;8,2>:uw 0x0080:uw
492 add.sat (8) r[a0.4,64]<2>:uw r[a0.4,64]<16;8,2>:uw 0x0080:uw
493
494 // restore pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4 registers
495 add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
496 add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
497
498 //Buffer 0
499 //Move Y to msg payload
500 mov (16) mubMSGPAYLOAD0(0,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
501 mov (16) mubMSGPAYLOAD0(0,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
502 mov (16) mubMSGPAYLOAD0(1,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
503 mov (16) mubMSGPAYLOAD0(1,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
504
505 //Move U to msg payload
506 mov (8) mubMSGPAYLOAD1(0,0)<2> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
507 mov (8) mubMSGPAYLOAD1(0,16)<2> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
508
509 //Move V to msg payload
510 mov (8) mubMSGPAYLOAD1(0,1)<2> r[a0.0, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
511 mov (8) mubMSGPAYLOAD1(0,17)<2> r[a0.0, 65]<32;8,4>:ub { NoDDChk }
512
513 add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
514
515 //Buffer 1
516 mov (16) mubMSGPAYLOAD0(2,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
517 mov (16) mubMSGPAYLOAD0(2,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
518 mov (16) mubMSGPAYLOAD0(3,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
519 mov (16) mubMSGPAYLOAD0(3,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
520
521 mov (8) mubMSGPAYLOAD1(1,0)<2> r[a0.6, 1]<32;8,4>:ub { NoDDClr }
522 mov (8) mubMSGPAYLOAD1(1,16)<2> r[a0.6, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
523
524 mov (8) mubMSGPAYLOAD1(1,1)<2> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
525 mov (8) mubMSGPAYLOAD1(1,17)<2> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
526
527 add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
528
529 //Buffer 2
530 mov (16) mubMSGPAYLOAD0(4,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
531 mov (16) mubMSGPAYLOAD0(4,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
532 mov (16) mubMSGPAYLOAD0(5,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
533 mov (16) mubMSGPAYLOAD0(5,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
534
535 mov (8) mubMSGPAYLOAD1(2,0)<2> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
536 mov (8) mubMSGPAYLOAD1(2,16)<2> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
537
538 mov (8) mubMSGPAYLOAD1(2,1)<2> r[a0.0, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
539 mov (8) mubMSGPAYLOAD1(2,17)<2> r[a0.0, 65]<32;8,4>:ub { NoDDChk }
540
541 //Buffer 3
542 mov (16) mubMSGPAYLOAD0(6,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
543 mov (16) mubMSGPAYLOAD0(6,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
544 mov (16) mubMSGPAYLOAD0(7,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
545 mov (16) mubMSGPAYLOAD0(7,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
546
547 mov (8) mubMSGPAYLOAD1(3,0)<2> r[a0.6, 1]<32;8,4>:ub { NoDDClr }
548 mov (8) mubMSGPAYLOAD1(3,16)<2> r[a0.6, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
549
550 mov (8) mubMSGPAYLOAD1(3,1)<2> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
551 mov (8) mubMSGPAYLOAD1(3,17)<2> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
552 //===========================================================================
553
554 send (1) null<1>:d r28 0xc 0x120A8018:ud
555 send (1) null<1>:d r37 0xc 0xA0A8019:ud