Debian Codegraph @

src/shaders/post_processing/gen7/Save_AVS_RGB.g4a

raw
1 /*
2 * Copyright 2000-2011 Intel Corporation All Rights Reserved
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the
6 * "Software"), to deal in the Software without restriction, including
7 * without limitation the rights to use, copy, modify, merge, publish,
8 * distribute, sub license, and/or sell copies of the Software, and to
9 * permit persons to whom the Software is furnished to do so, subject to
10 * the following conditions:
11 *
12 * The above copyright notice and this permission notice (including the
13 * next paragraph) shall be included in all copies or substantial portions
14 * of the Software.
15 *
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
17 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
18 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
19 * IN NO EVENT SHALL PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR
20 * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
21 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
22 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
23 *
24 * This file was originally licensed under the following license
25 *
26 * Licensed under the Apache License, Version 2.0 (the "License");
27 * you may not use this file except in compliance with the License.
28 * You may obtain a copy of the License at
29 *
30 * http://www.apache.org/licenses/LICENSE-2.0
31 *
32 * Unless required by applicable law or agreed to in writing, software
33 * distributed under the License is distributed on an "AS IS" BASIS,
34 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
35 * See the License for the specific language governing permissions and
36 * limitations under the License.
37 */
38 // 198 // Total instruction count
39 // 1 // Total kernel count
40
41
42
43 // Module name: common.inc
44 //
45 // Common header file for all Video-Processing kernels
46 //
47
48 .default_execution_size (16)
49 .default_register_type :ub
50
51 .reg_count_total 128
52 .reg_count_payload 7
53
54 //========== Common constants ==========
55
56
57 //========== Macros ==========
58
59
60 //Fast Jump, For more details see "Set_Layer_N.asm"
61
62
63 //========== Defines ====================
64
65 //========== Static Parameters (Common To All) ==========
66 //r1
67
68
69 //r2
70
71 // e.g. byte0 byte1 byte2
72 // YUYV 0 1 3
73 // YVYU 0 3 1
74
75 //Color Pipe (IECP) parameters
76
77
78 //ByteCopy
79
80
81 //r4
82
83 // e.g. byte0 byte1 byte2
84 // YUYV 0 1 3
85 // YVYU 0 3 1
86
87
88 //========== Inline parameters (Common To All) ===========
89
90
91 //============== Binding Index Table===========
92 //Common between DNDI and DNUV
93
94
95 //================= Common Message Descriptor =====
96 // Message descriptor for thread spawning
97 // Message Descriptors
98 // = 000 0001 (min message len 1 ) 0,0000 (resp len 0 -add later)
99 // 0000,0000,0000
100 // 0001(Spawn a root thread),0001 (Root thread spawn thread)
101 // = 0x02000011
102 // Thread Spawner Message Descriptor
103
104
105 // Message descriptor for atomic operation add
106 // Message Descriptors
107 // = 000 0110 (min message len 6 ) 0,0000 (resp len 0 -add later)
108 // 1(header present)001,10(typed atomic operation)0(return enabled)0(slot group, low 8 bits),0111 (AOP_Add)
109 // 0000,0000 (Binding table index, added later)
110 // = 0x02000011
111
112 // Atomic Operation Add Message Descriptor
113
114
115 // Message descriptor for dataport media write
116 // Message Descriptors
117 // = 000 0001 (min message len 1 - add later) 00000 (resp len 0)
118 // 1 (header present 1) 0 1010 (media block write) 000000
119 // 00000000 (binding table index - set later)
120 // = 0x020A8000
121
122
123 // Message Length defines
124
125
126 // Response Length defines
127
128
129 // Block Width and Height Size defines
130
131
132 // Extended Message Descriptors
133
134
135 // Common message descriptors:
136
137
138 //===================== Math Function Control ===================================
139
140
141 //============ Message Registers ===============
142 // buf4 starts from r28
143
144
145 //#define mMSGHDR_EOT r43 // Dummy Message Register for EOT
146
147
148 .declare mubMSGPAYLOAD Base=r30 ElementSize=1 SrcRegion=<16;16,1> Type=ub
149 .declare muwMSGPAYLOAD Base=r30 ElementSize=2 SrcRegion=<16;16,1> Type=uw
150 .declare mudMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=ud
151 .declare mfMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=f
152
153 //=================== End of thread instruction ===========================
154
155
156 //=====================Pointers Used=====================================
157
158
159 //=======================================================================
160
161
162 //r11-r17
163 // Define temp space for any usages
164
165
166 // Common Buffers
167
168
169 // temp space for rotation
170
171 .declare fROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
172
173 .declare udROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
174
175 .declare uwROBUF Base=r11.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
176
177 .declare ubROBUF Base=r11.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
178
179 .declare ub4ROBUF Base=r11.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
180
181
182 // End of common.inc
183
184
185 // Module name: Save_AVS_RGB.asm
186 //
187 // Save packed ARGB 444 frame data block of size 16x16
188 //
189 // To save 16x16 block (64x16 byte layout for ARGB8888) we need 4 send instructions with 16x16 in each
190 // -----------------
191 // | 0 | 1 | 2 | 3 |
192 // -----------------
193 // the 4 16x16 block send has been replaced by 16 32x2 sends to get better performance
194
195
196
197 // Module name: Save.inc
198
199
200
201
202 // Description: Includes all definitions explicit to Fast Composite.
203
204
205
206
207 // End of common.inc
208
209
210 //========== GRF partition ==========
211 // r0 header : r0 (1 GRF)
212 // Static parameters : r1 - r6 (6 GRFS)
213 // Inline parameters : r7 - r8 (2 GRFs)
214 // MSGSRC : r27 (1 GRF)
215 //===================================
216
217 //Interface:
218 //========== Static Parameters (Explicit To Fast Composite) ==========
219 //r1
220 //CSC Set 0
221
222
223 .declare udCSC_CURBE Base=r1.0 ElementSize=4 Type=ud
224
225 //Constant alpha
226
227
228 //r2
229
230
231 // Gen7 AVS WA
232
233
234 // WiDi Definitions
235
236
237 //Colorfill
238
239
240 // 0: 0-degree, 1: 90, 2: 180, 3: 270-degree, clockwise.
241
242 .declare ubCOLOR_PIXEL_VAL Base=r2.20 ElementSize=1 SrcRegion=<0;1,0> DstRegion=<1> Type=ub
243
244 //r3
245 //Normalised Ratio of Horizontal step size with main video for all layers
246
247
248 //Normalised Ratio of Horizontal step size with main video for all layers becomes
249 //Normalised Horizontal step size for all layers in VP_Setup.asm
250
251
252 //r4
253 //Normalised Vertical step size for all layers
254
255
256 //r5
257 //Normalised Vertical Frame Origin for all layers
258
259
260 //r6
261 //Normalised Horizontal Frame Origin for all layers
262
263
264 //========== Inline Parameters (Explicit To Fast Composite) ==========
265
266
267 //Main video Step X
268
269
270 //====================== Binding table (Explicit To Fast Composite)=========================================
271
272
273 //Used by Interlaced Scaling Kernels
274
275
276 //========== Sampler State Table Index (Explicit To Fast Composite)==========
277 //Sampler Index for AVS/IEF messages
278
279
280 //Sampler Index for SIMD16 sampler messages
281
282
283 //=============================================================================
284
285 .declare fBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
286 .declare fBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
287 .declare fBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
288 .declare fBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
289 .declare fBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
290 .declare fBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
291
292 .declare udBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
293 .declare udBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
294 .declare udBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
295 .declare udBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
296 .declare udBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
297 .declare udBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
298
299 .declare uwBUFFER_0 Base=r64.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
300 .declare uwBUFFER_1 Base=r80.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
301 .declare uwBUFFER_2 Base=r96.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
302 .declare uwBUFFER_3 Base=r112.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
303 .declare uwBUFFER_4 Base=r28.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
304 .declare uwBUFFER_5 Base=r46.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
305
306 .declare ubBUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
307 .declare ubBUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
308 .declare ubBUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
309 .declare ubBUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
310 .declare ubBUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
311 .declare ubBUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
312
313 .declare ub4BUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
314 .declare ub4BUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
315 .declare ub4BUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
316 .declare ub4BUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
317 .declare ub4BUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
318 .declare ub4BUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
319
320 //Pointer to mask reg
321
322
323 //r18
324
325
326 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
327 // NODDCLR, NODDCHK flags. -rT
328
329
330 .declare udCSC_COEFF_0 Base=r18.0 ElementSize=4 Type=ud // 1 GRF
331
332 //r19
333
334
335 .declare udCSC_COEFF_1 Base=r19.0 ElementSize=4 Type=ud // 1 GRF
336
337
338 //r20
339
340 .declare uwALPHA_MASK_REG_TEMP Base=r20.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
341
342 //r21
343
344 .declare uwALPHA_MASK_REG Base=r21.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
345
346 //r22
347
348
349 //Always keep Cannel Pointers and Offsets in same GRF, so that we can use
350 // NODDCLR, NODDCHK flags. -rT
351
352
353 //Keep fORIGIN_X_NLAS, fY_OFFSET_2ND_BLOCK, fSTEP_X_NLAS, pMSGDSC_COPY, ubCONST_ALPHA_COPY as
354 //sub registers of same GRF to enable using NODDCLR NODDCHK. -rT
355
356 //r23
357
358
359 //Lumakey
360
361
362 //r24
363
364
365 //r25
366
367
368 //r26
369
370
371 //defines to generate LABELS during compile time.
372
373
374 //Msg payload buffers; upto 4 full-size messages can be written
375
376
377 .declare mudMSGPAYLOAD0 Base=r29.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
378 .declare mudMSGPAYLOAD1 Base=r38.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
379 .declare mudMSGPAYLOAD2 Base=r47.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
380 .declare mudMSGPAYLOAD3 Base=r56.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
381
382 .declare muwMSGPAYLOAD0 Base=r29.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
383 .declare muwMSGPAYLOAD1 Base=r38.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
384 .declare muwMSGPAYLOAD2 Base=r47.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
385 .declare muwMSGPAYLOAD3 Base=r56.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
386
387 .declare mubMSGPAYLOAD0 Base=r29.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
388 .declare mubMSGPAYLOAD1 Base=r38.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
389 .declare mubMSGPAYLOAD2 Base=r47.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
390 .declare mubMSGPAYLOAD3 Base=r56.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
391 .declare mubMSGPAYLOAD4 Base=r32.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
392 .declare mubMSGPAYLOAD5 Base=r41.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
393 .declare mubMSGPAYLOAD6 Base=r50.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
394 .declare mubMSGPAYLOAD7 Base=r59.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
395
396
397 // the r17 register (nTEMP0) is originally defined from "Common.inc"
398 // instead of re-defining a nTEMP0 here, we use "SAVE_RGB" suffix for its naming
399
400 .declare uwTemp0 Base=r17.0 ElementSize=2 Type=uw
401
402
403 //_SAVE_INC_
404
405
406 // At the save module we have all 8 address sub-registers available.
407 // So we will use PING-PONG type of scheme to save the data using
408 // pointers pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4. This will help
409 // reduce dependency. - rT
410
411 // channel switching based on bit 0 of uWRGB_BGR_CH_SWITCH
412
413 // if channel swap?
414 and.nz.f0.0 null<1>:w r2.3<0;1,0>:uw 0x01:w
415
416 //wBUFF_CHNL_PTR points to either buffer 0 or buffer 4.
417 //Add appropriate offsets to get pointers for all buffers (1,2,3 or 5).
418 //Offsets are zero for buffer 0 and buffer 4.
419 add (4) a0.0:uw r22.0<4;4,1>:w 0:uw
420
421 // pointer swap
422 (f0.0) mov (1) uwTemp0<1> a0.0:uw
423 (f0.0) mov (1) a0.0:uw a0.2:uw
424 (f0.0) mov (1) a0.2:uw uwTemp0<0;1,0>
425
426 shl (1) r27.0<1>:d r9.0<0;1,0>:w 2:w { NoDDClr } // H. block origin need to be quadrupled
427 mov (1) r27.1<1>:d r9.1<0;1,0>:w { NoDDClr, NoDDChk } // Block origin (1st quadrant)
428 mov (1) r27.2<1>:ud 0x1001F:ud { NoDDChk } // Block width and height (32x2)
429
430 add (4) a0.4:uw a0.0<4;4,1>:w r22.8<0;2,1>:w
431
432 mov (8) r28<1>:ud r27<8;8,1>:ud
433 mov (8) r37<1>:ud r27<8;8,1>:ud
434 mov (8) r46<1>:ud r27<8;8,1>:ud
435 mov (8) r55<1>:ud r27<8;8,1>:ud
436
437 mov (8) r31<1>:ud r27<8;8,1>:ud
438 mov (8) r40<1>:ud r27<8;8,1>:ud
439 mov (8) r49<1>:ud r27<8;8,1>:ud
440 mov (8) r58<1>:ud r27<8;8,1>:ud
441
442
443 //for BUFFER 0
444 add (1) r37.1<1>:d r27.1<0;1,0>:d 2:d
445 add (1) r46.0<1>:d r27.0<0;1,0>:d 32:d
446 add (1) r55.0<1>:d r27.0<0;1,0>:d 32:d
447 add (1) r55.1<1>:d r27.1<0;1,0>:d 2:d
448
449 // for BUFFER 1
450 add (1) r31.1<1>:d r27.1<0;1,0>:d 4:d
451 add (1) r40.1<1>:d r27.1<0;1,0>:d 6:d
452 add (1) r49.0<1>:d r27.0<0;1,0>:d 32:d
453 add (1) r49.1<1>:d r27.1<0;1,0>:d 4:d
454 add (1) r58.0<1>:d r27.0<0;1,0>:d 32:d
455 add (1) r58.1<1>:d r27.1<0;1,0>:d 6:d
456 // write Buf_0 to 1st quarter of four horizontal output blocks
457
458 // Please note the scattered order of NODDCLR, NODDCHK flags. Since the sub-registers
459 // of destination reg are not updated at one place and hence even flags are scattered. -rT
460
461 mov (8) mubMSGPAYLOAD0(0, 0)<4> r[a0.2, 1]<16;8,2> { NoDDClr }
462 mov (8) mubMSGPAYLOAD0(0, 1)<4> r[a0.1, 1]<16;8,2> { NoDDClr, NoDDChk }
463 mov (8) mubMSGPAYLOAD0(0, 2)<4> r[a0.0, 1]<16;8,2> { NoDDClr, NoDDChk }
464 mov (8) mubMSGPAYLOAD0(0, 3)<4> r2.31:ub { NoDDChk }
465
466 mov (8) mubMSGPAYLOAD0(1, 0)<4> r[a0.2, 33]<16;8,2> { NoDDClr }
467 mov (8) mubMSGPAYLOAD0(1, 1)<4> r[a0.1, 33]<16;8,2> { NoDDClr, NoDDChk }
468 mov (8) mubMSGPAYLOAD0(1, 2)<4> r[a0.0, 33]<16;8,2> { NoDDClr, NoDDChk }
469 mov (8) mubMSGPAYLOAD0(1, 3)<4> r2.31:ub { NoDDChk }
470
471 mov (8) mubMSGPAYLOAD1(0, 0)<4> r[a0.6, 1]<16;8,2> { NoDDClr }
472 mov (8) mubMSGPAYLOAD1(0, 1)<4> r[a0.5, 1]<16;8,2> { NoDDClr, NoDDChk }
473 mov (8) mubMSGPAYLOAD1(0, 2)<4> r[a0.4, 1]<16;8,2> { NoDDClr, NoDDChk }
474 mov (8) mubMSGPAYLOAD1(0, 3)<4> r2.31:ub { NoDDChk }
475
476 mov (8) mubMSGPAYLOAD1(1, 0)<4> r[a0.6, 33]<16;8,2> { NoDDClr }
477 mov (8) mubMSGPAYLOAD1(1, 1)<4> r[a0.5, 33]<16;8,2> { NoDDClr, NoDDChk }
478 mov (8) mubMSGPAYLOAD1(1, 2)<4> r[a0.4, 33]<16;8,2> { NoDDClr, NoDDChk }
479 mov (8) mubMSGPAYLOAD1(1, 3)<4> r2.31:ub { NoDDChk }
480
481 mov (8) mubMSGPAYLOAD2(0, 0)<4> r[a0.2, 17]<16;8,2> { NoDDClr }
482 mov (8) mubMSGPAYLOAD2(0, 1)<4> r[a0.1, 17]<16;8,2> { NoDDClr, NoDDChk }
483 mov (8) mubMSGPAYLOAD2(0, 2)<4> r[a0.0, 17]<16;8,2> { NoDDClr, NoDDChk }
484 mov (8) mubMSGPAYLOAD2(0, 3)<4> r2.31:ub { NoDDChk }
485
486 mov (8) mubMSGPAYLOAD2(1, 0)<4> r[a0.2, 49]<16;8,2> { NoDDClr }
487 mov (8) mubMSGPAYLOAD2(1, 1)<4> r[a0.1, 49]<16;8,2> { NoDDClr, NoDDChk }
488 mov (8) mubMSGPAYLOAD2(1, 2)<4> r[a0.0, 49]<16;8,2> { NoDDClr, NoDDChk }
489 mov (8) mubMSGPAYLOAD2(1, 3)<4> r2.31:ub { NoDDChk }
490
491 mov (8) mubMSGPAYLOAD3(0, 0)<4> r[a0.6, 17]<16;8,2> { NoDDClr }
492 mov (8) mubMSGPAYLOAD3(0, 1)<4> r[a0.5, 17]<16;8,2> { NoDDClr, NoDDChk }
493 mov (8) mubMSGPAYLOAD3(0, 2)<4> r[a0.4, 17]<16;8,2> { NoDDClr, NoDDChk }
494 mov (8) mubMSGPAYLOAD3(0, 3)<4> r2.31:ub { NoDDChk }
495
496 mov (8) mubMSGPAYLOAD3(1, 0)<4> r[a0.6, 49]<16;8,2> { NoDDClr }
497 mov (8) mubMSGPAYLOAD3(1, 1)<4> r[a0.5, 49]<16;8,2> { NoDDClr, NoDDChk }
498 mov (8) mubMSGPAYLOAD3(1, 2)<4> r[a0.4, 49]<16;8,2> { NoDDClr, NoDDChk }
499 mov (8) mubMSGPAYLOAD3(1, 3)<4> r2.31:ub { NoDDChk }
500
501 // write Buf_1 to 2nd quarter of four horizontal output blocks
502 add (4) a0.0:uw r22.0<4;4,1>:w 512:uw
503
504 // pointer swap
505 (f0.0) mov (1) uwTemp0<1> a0.0:uw
506 (f0.0) mov (1) a0.0:uw a0.2:uw
507 (f0.0) mov (1) a0.2:uw uwTemp0<0;1,0>
508
509 add (4) a0.4:uw a0.0<4;4,1>:w r22.8<0;2,1>:w
510
511 mov (8) mubMSGPAYLOAD4(0, 0)<4> r[a0.2, 1]<16;8,2> { NoDDClr }
512 mov (8) mubMSGPAYLOAD4(0, 1)<4> r[a0.1, 1]<16;8,2> { NoDDClr, NoDDChk }
513 mov (8) mubMSGPAYLOAD4(0, 2)<4> r[a0.0, 1]<16;8,2> { NoDDClr, NoDDChk }
514 mov (8) mubMSGPAYLOAD4(0, 3)<4> r2.31:ub { NoDDChk }
515
516 mov (8) mubMSGPAYLOAD4(1, 0)<4> r[a0.2, 33]<16;8,2> { NoDDClr }
517 mov (8) mubMSGPAYLOAD4(1, 1)<4> r[a0.1, 33]<16;8,2> { NoDDClr, NoDDChk }
518 mov (8) mubMSGPAYLOAD4(1, 2)<4> r[a0.0, 33]<16;8,2> { NoDDClr, NoDDChk }
519 mov (8) mubMSGPAYLOAD4(1, 3)<4> r2.31:ub { NoDDChk }
520
521 mov (8) mubMSGPAYLOAD5(0, 0)<4> r[a0.6, 1]<16;8,2> { NoDDClr }
522 mov (8) mubMSGPAYLOAD5(0, 1)<4> r[a0.5, 1]<16;8,2> { NoDDClr, NoDDChk }
523 mov (8) mubMSGPAYLOAD5(0, 2)<4> r[a0.4, 1]<16;8,2> { NoDDClr, NoDDChk }
524 mov (8) mubMSGPAYLOAD5(0, 3)<4> r2.31:ub { NoDDChk }
525
526 mov (8) mubMSGPAYLOAD5(1, 0)<4> r[a0.6, 33]<16;8,2> { NoDDClr }
527 mov (8) mubMSGPAYLOAD5(1, 1)<4> r[a0.5, 33]<16;8,2> { NoDDClr, NoDDChk }
528 mov (8) mubMSGPAYLOAD5(1, 2)<4> r[a0.4, 33]<16;8,2> { NoDDClr, NoDDChk }
529 mov (8) mubMSGPAYLOAD5(1, 3)<4> r2.31:ub { NoDDChk }
530
531 mov (8) mubMSGPAYLOAD6(0, 0)<4> r[a0.2, 17]<16;8,2> { NoDDClr }
532 mov (8) mubMSGPAYLOAD6(0, 1)<4> r[a0.1, 17]<16;8,2> { NoDDClr, NoDDChk }
533 mov (8) mubMSGPAYLOAD6(0, 2)<4> r[a0.0, 17]<16;8,2> { NoDDClr, NoDDChk }
534 mov (8) mubMSGPAYLOAD6(0, 3)<4> r2.31:ub { NoDDChk }
535
536 mov (8) mubMSGPAYLOAD6(1, 0)<4> r[a0.2, 49]<16;8,2> { NoDDClr }
537 mov (8) mubMSGPAYLOAD6(1, 1)<4> r[a0.1, 49]<16;8,2> { NoDDClr, NoDDChk }
538 mov (8) mubMSGPAYLOAD6(1, 2)<4> r[a0.0, 49]<16;8,2> { NoDDClr, NoDDChk }
539 mov (8) mubMSGPAYLOAD6(1, 3)<4> r2.31:ub { NoDDChk }
540
541 mov (8) mubMSGPAYLOAD7(0, 0)<4> r[a0.6, 17]<16;8,2> { NoDDClr }
542 mov (8) mubMSGPAYLOAD7(0, 1)<4> r[a0.5, 17]<16;8,2> { NoDDClr, NoDDChk }
543 mov (8) mubMSGPAYLOAD7(0, 2)<4> r[a0.4, 17]<16;8,2> { NoDDClr, NoDDChk }
544 mov (8) mubMSGPAYLOAD7(0, 3)<4> r2.31:ub { NoDDChk }
545
546 mov (8) mubMSGPAYLOAD7(1, 0)<4> r[a0.6, 49]<16;8,2> { NoDDClr }
547 mov (8) mubMSGPAYLOAD7(1, 1)<4> r[a0.5, 49]<16;8,2> { NoDDClr, NoDDChk }
548 mov (8) mubMSGPAYLOAD7(1, 2)<4> r[a0.4, 49]<16;8,2> { NoDDClr, NoDDChk }
549 mov (8) mubMSGPAYLOAD7(1, 3)<4> r2.31:ub { NoDDChk }
550
551 // send buffer 0 and buffer 1
552 send (1) null<1>:d r28 0x5 0x60A8018:ud
553 send (1) null<1>:d r37 0x5 0x60A8018:ud
554 send (1) null<1>:d r46 0x5 0x60A8018:ud
555 send (1) null<1>:d r55 0x5 0x60A8018:ud
556
557 send (1) null<1>:d r31 0x5 0x60A8018:ud
558 send (1) null<1>:d r40 0x5 0x60A8018:ud
559 send (1) null<1>:d r49 0x5 0x60A8018:ud
560 send (1) null<1>:d r58 0x5 0x60A8018:ud
561
562 //==========
563 //prepare headers
564 //for BUFFER 2
565 add (1) r28.1<1>:d r27.1<0;1,0>:d 8:d
566 add (1) r37.1<1>:d r27.1<0;1,0>:d 10:d
567 add (1) r46.0<1>:d r27.0<0;1,0>:d 32:d
568 add (1) r46.1<1>:d r27.1<0;1,0>:d 8:d
569 add (1) r55.0<1>:d r27.0<0;1,0>:d 32:d
570 add (1) r55.1<1>:d r27.1<0;1,0>:d 10:d
571 // for BUFFER 3
572 add (1) r31.1<1>:d r27.1<0;1,0>:d 12:d
573 add (1) r40.1<1>:d r27.1<0;1,0>:d 14:d
574 add (1) r49.0<1>:d r27.0<0;1,0>:d 32:d
575 add (1) r49.1<1>:d r27.1<0;1,0>:d 12:d
576 add (1) r58.0<1>:d r27.0<0;1,0>:d 32:d
577 add (1) r58.1<1>:d r27.1<0;1,0>:d 14:d
578
579 //===========
580
581 // write Buf_2 to 3rd quarter of four horizontal output blocks
582 add (4) a0.0:uw r22.0<4;4,1>:w 1024:uw
583
584 // pointer swap
585 (f0.0) mov (1) uwTemp0<1> a0.0:uw
586 (f0.0) mov (1) a0.0:uw a0.2:uw
587 (f0.0) mov (1) a0.2:uw uwTemp0<0;1,0>
588
589 add (4) a0.4:uw a0.0<4;4,1>:w r22.8<0;2,1>:w
590
591 mov (8) mubMSGPAYLOAD0(0, 0)<4> r[a0.2, 1]<16;8,2> { NoDDClr }
592 mov (8) mubMSGPAYLOAD0(0, 1)<4> r[a0.1, 1]<16;8,2> { NoDDClr, NoDDChk }
593 mov (8) mubMSGPAYLOAD0(0, 2)<4> r[a0.0, 1]<16;8,2> { NoDDClr, NoDDChk }
594 mov (8) mubMSGPAYLOAD0(0, 3)<4> r2.31:ub { NoDDChk }
595
596 mov (8) mubMSGPAYLOAD0(1, 0)<4> r[a0.2, 33]<16;8,2> { NoDDClr }
597 mov (8) mubMSGPAYLOAD0(1, 1)<4> r[a0.1, 33]<16;8,2> { NoDDClr, NoDDChk }
598 mov (8) mubMSGPAYLOAD0(1, 2)<4> r[a0.0, 33]<16;8,2> { NoDDClr, NoDDChk }
599 mov (8) mubMSGPAYLOAD0(1, 3)<4> r2.31:ub { NoDDChk }
600
601 mov (8) mubMSGPAYLOAD1(0, 0)<4> r[a0.6, 1]<16;8,2> { NoDDClr }
602 mov (8) mubMSGPAYLOAD1(0, 1)<4> r[a0.5, 1]<16;8,2> { NoDDClr, NoDDChk }
603 mov (8) mubMSGPAYLOAD1(0, 2)<4> r[a0.4, 1]<16;8,2> { NoDDClr, NoDDChk }
604 mov (8) mubMSGPAYLOAD1(0, 3)<4> r2.31:ub { NoDDChk }
605
606 mov (8) mubMSGPAYLOAD1(1, 0)<4> r[a0.6, 33]<16;8,2> { NoDDClr }
607 mov (8) mubMSGPAYLOAD1(1, 1)<4> r[a0.5, 33]<16;8,2> { NoDDClr, NoDDChk }
608 mov (8) mubMSGPAYLOAD1(1, 2)<4> r[a0.4, 33]<16;8,2> { NoDDClr, NoDDChk }
609 mov (8) mubMSGPAYLOAD1(1, 3)<4> r2.31:ub { NoDDChk }
610
611 mov (8) mubMSGPAYLOAD2(0, 0)<4> r[a0.2, 17]<16;8,2> { NoDDClr }
612 mov (8) mubMSGPAYLOAD2(0, 1)<4> r[a0.1, 17]<16;8,2> { NoDDClr, NoDDChk }
613 mov (8) mubMSGPAYLOAD2(0, 2)<4> r[a0.0, 17]<16;8,2> { NoDDClr, NoDDChk }
614 mov (8) mubMSGPAYLOAD2(0, 3)<4> r2.31:ub { NoDDChk }
615
616 mov (8) mubMSGPAYLOAD2(1, 0)<4> r[a0.2, 49]<16;8,2> { NoDDClr }
617 mov (8) mubMSGPAYLOAD2(1, 1)<4> r[a0.1, 49]<16;8,2> { NoDDClr, NoDDChk }
618 mov (8) mubMSGPAYLOAD2(1, 2)<4> r[a0.0, 49]<16;8,2> { NoDDClr, NoDDChk }
619 mov (8) mubMSGPAYLOAD2(1, 3)<4> r2.31:ub { NoDDChk }
620
621 mov (8) mubMSGPAYLOAD3(0, 0)<4> r[a0.6, 17]<16;8,2> { NoDDClr }
622 mov (8) mubMSGPAYLOAD3(0, 1)<4> r[a0.5, 17]<16;8,2> { NoDDClr, NoDDChk }
623 mov (8) mubMSGPAYLOAD3(0, 2)<4> r[a0.4, 17]<16;8,2> { NoDDClr, NoDDChk }
624 mov (8) mubMSGPAYLOAD3(0, 3)<4> r2.31:ub { NoDDChk }
625
626 mov (8) mubMSGPAYLOAD3(1, 0)<4> r[a0.6, 49]<16;8,2> { NoDDClr }
627 mov (8) mubMSGPAYLOAD3(1, 1)<4> r[a0.5, 49]<16;8,2> { NoDDClr, NoDDChk }
628 mov (8) mubMSGPAYLOAD3(1, 2)<4> r[a0.4, 49]<16;8,2> { NoDDClr, NoDDChk }
629 mov (8) mubMSGPAYLOAD3(1, 3)<4> r2.31:ub { NoDDChk }
630
631 // write Buf_3 to 4th quarter of four horizontal output blocks
632 add (4) a0.0:uw r22.0<4;4,1>:w 1536:uw
633
634 // pointer swap
635 (f0.0) mov (1) uwTemp0<1> a0.0:uw
636 (f0.0) mov (1) a0.0:uw a0.2:uw
637 (f0.0) mov (1) a0.2:uw uwTemp0<0;1,0>
638
639 add (4) a0.4:uw a0.0<4;4,1>:w r22.8<0;2,1>:w
640
641 mov (8) mubMSGPAYLOAD4(0, 0)<4> r[a0.2, 1]<16;8,2> { NoDDClr }
642 mov (8) mubMSGPAYLOAD4(0, 1)<4> r[a0.1, 1]<16;8,2> { NoDDClr, NoDDChk }
643 mov (8) mubMSGPAYLOAD4(0, 2)<4> r[a0.0, 1]<16;8,2> { NoDDClr, NoDDChk }
644 mov (8) mubMSGPAYLOAD4(0, 3)<4> r2.31:ub { NoDDChk }
645
646 mov (8) mubMSGPAYLOAD4(1, 0)<4> r[a0.2, 33]<16;8,2> { NoDDClr }
647 mov (8) mubMSGPAYLOAD4(1, 1)<4> r[a0.1, 33]<16;8,2> { NoDDClr, NoDDChk }
648 mov (8) mubMSGPAYLOAD4(1, 2)<4> r[a0.0, 33]<16;8,2> { NoDDClr, NoDDChk }
649 mov (8) mubMSGPAYLOAD4(1, 3)<4> r2.31:ub { NoDDChk }
650
651 mov (8) mubMSGPAYLOAD5(0, 0)<4> r[a0.6, 1]<16;8,2> { NoDDClr }
652 mov (8) mubMSGPAYLOAD5(0, 1)<4> r[a0.5, 1]<16;8,2> { NoDDClr, NoDDChk }
653 mov (8) mubMSGPAYLOAD5(0, 2)<4> r[a0.4, 1]<16;8,2> { NoDDClr, NoDDChk }
654 mov (8) mubMSGPAYLOAD5(0, 3)<4> r2.31:ub { NoDDChk }
655
656 mov (8) mubMSGPAYLOAD5(1, 0)<4> r[a0.6, 33]<16;8,2> { NoDDClr }
657 mov (8) mubMSGPAYLOAD5(1, 1)<4> r[a0.5, 33]<16;8,2> { NoDDClr, NoDDChk }
658 mov (8) mubMSGPAYLOAD5(1, 2)<4> r[a0.4, 33]<16;8,2> { NoDDClr, NoDDChk }
659 mov (8) mubMSGPAYLOAD5(1, 3)<4> r2.31:ub { NoDDChk }
660
661 mov (8) mubMSGPAYLOAD6(0, 0)<4> r[a0.2, 17]<16;8,2> { NoDDClr }
662 mov (8) mubMSGPAYLOAD6(0, 1)<4> r[a0.1, 17]<16;8,2> { NoDDClr, NoDDChk }
663 mov (8) mubMSGPAYLOAD6(0, 2)<4> r[a0.0, 17]<16;8,2> { NoDDClr, NoDDChk }
664 mov (8) mubMSGPAYLOAD6(0, 3)<4> r2.31:ub { NoDDChk }
665
666 mov (8) mubMSGPAYLOAD6(1, 0)<4> r[a0.2, 49]<16;8,2> { NoDDClr }
667 mov (8) mubMSGPAYLOAD6(1, 1)<4> r[a0.1, 49]<16;8,2> { NoDDClr, NoDDChk }
668 mov (8) mubMSGPAYLOAD6(1, 2)<4> r[a0.0, 49]<16;8,2> { NoDDClr, NoDDChk }
669 mov (8) mubMSGPAYLOAD6(1, 3)<4> r2.31:ub { NoDDChk }
670
671 mov (8) mubMSGPAYLOAD7(0, 0)<4> r[a0.6, 17]<16;8,2> { NoDDClr }
672 mov (8) mubMSGPAYLOAD7(0, 1)<4> r[a0.5, 17]<16;8,2> { NoDDClr, NoDDChk }
673 mov (8) mubMSGPAYLOAD7(0, 2)<4> r[a0.4, 17]<16;8,2> { NoDDClr, NoDDChk }
674 mov (8) mubMSGPAYLOAD7(0, 3)<4> r2.31:ub { NoDDChk }
675
676 mov (8) mubMSGPAYLOAD7(1, 0)<4> r[a0.6, 49]<16;8,2> { NoDDClr }
677 mov (8) mubMSGPAYLOAD7(1, 1)<4> r[a0.5, 49]<16;8,2> { NoDDClr, NoDDChk }
678 mov (8) mubMSGPAYLOAD7(1, 2)<4> r[a0.4, 49]<16;8,2> { NoDDClr, NoDDChk }
679 mov (8) mubMSGPAYLOAD7(1, 3)<4> r2.31:ub { NoDDChk }
680
681 // send buffer 2 and buffer 3
682 send (1) null<1>:d r28 0x5 0x60A8018:ud
683 send (1) null<1>:d r37 0x5 0x60A8018:ud
684 send (1) null<1>:d r46 0x5 0x60A8018:ud
685 send (1) null<1>:d r55 0x5 0x60A8018:ud
686
687 send (1) null<1>:d r31 0x5 0x60A8018:ud
688 send (1) null<1>:d r40 0x5 0x60A8018:ud
689 send (1) null<1>:d r49 0x5 0x60A8018:ud
690 send (1) null<1>:d r58 0x5 0x60A8018:ud