src/gallium/docs/source/tgsi.rst

   1 TGSI
   2 ====
   3
   4 TGSI, Tungsten Graphics Shader Instructions, is an intermediate language
   5 for describing shaders. Since Gallium is inherently shaderful, shaders are
   6 an important part of the API. TGSI is the only intermediate representation
   7 used by all drivers.
   8
   9 From GL_NV_vertex_program
  10 -------------------------
  11
  12
  13 ARL - Address Register Load
  14
  15 .. math::
  16
  17   dst.x = \lfloor src.x\rfloor
  18
  19   dst.y = \lfloor src.y\rfloor
  20
  21   dst.z = \lfloor src.z\rfloor
  22
  23   dst.w = \lfloor src.w\rfloor
  24
  25
  26 MOV - Move
  27
  28 .. math::
  29
  30   dst.x = src.x
  31
  32   dst.y = src.y
  33
  34   dst.z = src.z
  35
  36   dst.w = src.w
  37
  38
  39 LIT - Light Coefficients
  40
  41 .. math::
  42
  43   dst.x = 1.0
  44
  45   dst.y = max(src.x, 0.0)
  46
  47   dst.z = (src.x > 0.0) ? max(src.y, 0.0)^{clamp(src.w, -128.0, 128.0))} : 0.0
  48
  49   dst.w = 1.0
  50
  51
  52 RCP - Reciprocal
  53
  54 .. math::
  55
  56   dst.x = 1.0 / src.x
  57
  58   dst.y = 1.0 / src.x
  59
  60   dst.z = 1.0 / src.x
  61
  62   dst.w = 1.0 / src.x
  63
  64
  65 RSQ - Reciprocal Square Root
  66
  67 .. math::
  68
  69   dst.x = 1.0 / \sqrt{abs(src.x)}
  70
  71   dst.y = 1.0 / \sqrt{abs(src.x)}
  72
  73   dst.z = 1.0 / \sqrt{abs(src.x)}
  74
  75   dst.w = 1.0 / \sqrt{abs(src.x)}
  76
  77
  78 EXP - Approximate Exponential Base 2
  79
  80 .. math::
  81
  82   dst.x = 2^{\lfloor src.x\rfloor}
  83
  84   dst.y = src.x - \lfloor src.x\rfloor
  85
  86   dst.z = 2^{src.x}
  87
  88   dst.w = 1.0
  89
  90
  91 LOG - Approximate Logarithm Base 2
  92
  93 .. math::
  94
  95   dst.x = \lfloor lg2(abs(src.x)))\rfloor
  96
  97   dst.y = abs(src.x) / 2^{\lfloor lg2(abs(src.x))\rfloor}
  98
  99   dst.z = lg2(abs(src.x))
 100
 101   dst.w = 1.0
 102
 103
 104 MUL - Multiply
 105
 106 .. math::
 107
 108   dst.x = src0.x * src1.x
 109
 110   dst.y = src0.y * src1.y
 111
 112   dst.z = src0.z * src1.z
 113
 114   dst.w = src0.w * src1.w
 115
 116
 117 ADD - Add
 118
 119 .. math::
 120
 121   dst.x = src0.x + src1.x
 122
 123   dst.y = src0.y + src1.y
 124
 125   dst.z = src0.z + src1.z
 126
 127   dst.w = src0.w + src1.w
 128
 129
 130 DP3 - 3-component Dot Product
 131
 132 .. math::
 133
 134   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
 135
 136   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
 137
 138   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
 139
 140   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
 141
 142
 143 DP4 - 4-component Dot Product
 144
 145 .. math::
 146
 147   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 148
 149   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 150
 151   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 152
 153   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 154
 155
 156 DST - Distance Vector
 157
 158 .. math::
 159
 160   dst.x = 1.0
 161
 162   dst.y = src0.y * src1.y
 163
 164   dst.z = src0.z
 165
 166   dst.w = src1.w
 167
 168
 169 MIN - Minimum
 170
 171 .. math::
 172
 173   dst.x = min(src0.x, src1.x)
 174
 175   dst.y = min(src0.y, src1.y)
 176
 177   dst.z = min(src0.z, src1.z)
 178
 179   dst.w = min(src0.w, src1.w)
 180
 181
 182 MAX - Maximum
 183
 184 .. math::
 185
 186   dst.x = max(src0.x, src1.x)
 187
 188   dst.y = max(src0.y, src1.y)
 189
 190   dst.z = max(src0.z, src1.z)
 191
 192   dst.w = max(src0.w, src1.w)
 193
 194
 195 SLT - Set On Less Than
 196
 197 .. math::
 198
 199   dst.x = (src0.x < src1.x) ? 1.0 : 0.0
 200
 201   dst.y = (src0.y < src1.y) ? 1.0 : 0.0
 202
 203   dst.z = (src0.z < src1.z) ? 1.0 : 0.0
 204
 205   dst.w = (src0.w < src1.w) ? 1.0 : 0.0
 206
 207
 208 SGE - Set On Greater Equal Than
 209
 210 .. math::
 211
 212   dst.x = (src0.x >= src1.x) ? 1.0 : 0.0
 213
 214   dst.y = (src0.y >= src1.y) ? 1.0 : 0.0
 215
 216   dst.z = (src0.z >= src1.z) ? 1.0 : 0.0
 217
 218   dst.w = (src0.w >= src1.w) ? 1.0 : 0.0
 219
 220
 221 MAD - Multiply And Add
 222
 223 .. math::
 224
 225   dst.x = src0.x * src1.x + src2.x
 226
 227   dst.y = src0.y * src1.y + src2.y
 228
 229   dst.z = src0.z * src1.z + src2.z
 230
 231   dst.w = src0.w * src1.w + src2.w
 232
 233
 234 SUB - Subtract
 235
 236 .. math::
 237
 238   dst.x = src0.x - src1.x
 239
 240   dst.y = src0.y - src1.y
 241
 242   dst.z = src0.z - src1.z
 243
 244   dst.w = src0.w - src1.w
 245
 246
 247 LRP - Linear Interpolate
 248
 249 .. math::
 250
 251   dst.x = src0.x * (src1.x - src2.x) + src2.x
 252
 253   dst.y = src0.y * (src1.y - src2.y) + src2.y
 254
 255   dst.z = src0.z * (src1.z - src2.z) + src2.z
 256
 257   dst.w = src0.w * (src1.w - src2.w) + src2.w
 258
 259
 260 CND - Condition
 261
 262 .. math::
 263
 264   dst.x = (src2.x > 0.5) ? src0.x : src1.x
 265
 266   dst.y = (src2.y > 0.5) ? src0.y : src1.y
 267
 268   dst.z = (src2.z > 0.5) ? src0.z : src1.z
 269
 270   dst.w = (src2.w > 0.5) ? src0.w : src1.w
 271
 272
 273 DP2A - 2-component Dot Product And Add
 274
 275 .. math::
 276
 277   dst.x = src0.x * src1.x + src0.y * src1.y + src2.x
 278
 279   dst.y = src0.x * src1.x + src0.y * src1.y + src2.x
 280
 281   dst.z = src0.x * src1.x + src0.y * src1.y + src2.x
 282
 283   dst.w = src0.x * src1.x + src0.y * src1.y + src2.x
 284
 285
 286 FRAC - Fraction
 287
 288 .. math::
 289
 290   dst.x = src.x - \lfloor src.x\rfloor
 291
 292   dst.y = src.y - \lfloor src.y\rfloor
 293
 294   dst.z = src.z - \lfloor src.z\rfloor
 295
 296   dst.w = src.w - \lfloor src.w\rfloor
 297
 298
 299 CLAMP - Clamp
 300
 301 .. math::
 302
 303   dst.x = clamp(src0.x, src1.x, src2.x)
 304   dst.y = clamp(src0.y, src1.y, src2.y)
 305   dst.z = clamp(src0.z, src1.z, src2.z)
 306   dst.w = clamp(src0.w, src1.w, src2.w)
 307
 308
 309 FLR - Floor
 310
 311 This is identical to ARL.
 312
 313 .. math::
 314
 315   dst.x = \lfloor src.x\rfloor
 316
 317   dst.y = \lfloor src.y\rfloor
 318
 319   dst.z = \lfloor src.z\rfloor
 320
 321   dst.w = \lfloor src.w\rfloor
 322
 323
 324 1.3.9  ROUND - Round
 325
 326 .. math::
 327
 328   dst.x = round(src.x)
 329   dst.y = round(src.y)
 330   dst.z = round(src.z)
 331   dst.w = round(src.w)
 332
 333
 334 EX2 - Exponential Base 2
 335
 336 .. math::
 337
 338   dst.x = 2^{src.x}
 339
 340   dst.y = 2^{src.x}
 341
 342   dst.z = 2^{src.x}
 343
 344   dst.w = 2^{src.x}
 345
 346
 347 1.3.11  LG2 - Logarithm Base 2
 348
 349 .. math::
 350
 351   dst.x = lg2(src.x)
 352   dst.y = lg2(src.x)
 353   dst.z = lg2(src.x)
 354   dst.w = lg2(src.x)
 355
 356
 357 POW - Power
 358
 359 .. math::
 360
 361   dst.x = src0.x^{src1.x}
 362
 363   dst.y = src0.x^{src1.x}
 364
 365   dst.z = src0.x^{src1.x}
 366
 367   dst.w = src0.x^{src1.x}
 368
 369 1.3.15  XPD - Cross Product
 370
 371 .. math::
 372
 373   dst.x = src0.y * src1.z - src1.y * src0.z
 374   dst.y = src0.z * src1.x - src1.z * src0.x
 375   dst.z = src0.x * src1.y - src1.x * src0.y
 376   dst.w = 1.0
 377
 378
 379 1.4.1  ABS - Absolute
 380
 381 .. math::
 382
 383   dst.x = abs(src.x)
 384   dst.y = abs(src.y)
 385   dst.z = abs(src.z)
 386   dst.w = abs(src.w)
 387
 388
 389 1.4.2  RCC - Reciprocal Clamped
 390
 391 .. math::
 392
 393   dst.x = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
 394   dst.y = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
 395   dst.z = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
 396   dst.w = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
 397
 398
 399 1.4.3  DPH - Homogeneous Dot Product
 400
 401 .. math::
 402
 403   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 404   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 405   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 406   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 407
 408
 409 COS - Cosine
 410
 411 .. math::
 412
 413   dst.x = \cos{src.x}
 414
 415   dst.y = \cos{src.x}
 416
 417   dst.z = \cos{src.x}
 418
 419   dst.w = \cos{src.w}
 420
 421
 422 1.5.2  DDX - Derivative Relative To X
 423
 424 .. math::
 425
 426   dst.x = partialx(src.x)
 427   dst.y = partialx(src.y)
 428   dst.z = partialx(src.z)
 429   dst.w = partialx(src.w)
 430
 431
 432 1.5.3  DDY - Derivative Relative To Y
 433
 434 .. math::
 435
 436   dst.x = partialy(src.x)
 437   dst.y = partialy(src.y)
 438   dst.z = partialy(src.z)
 439   dst.w = partialy(src.w)
 440
 441
 442 1.5.7  KILP - Predicated Discard
 443
 444 .. math::
 445
 446   discard
 447
 448
 449 1.5.10  PK2H - Pack Two 16-bit Floats
 450
 451   TBD
 452
 453
 454 1.5.11  PK2US - Pack Two Unsigned 16-bit Scalars
 455
 456   TBD
 457
 458
 459 1.5.12  PK4B - Pack Four Signed 8-bit Scalars
 460
 461   TBD
 462
 463
 464 1.5.13  PK4UB - Pack Four Unsigned 8-bit Scalars
 465
 466   TBD
 467
 468
 469 1.5.15  RFL - Reflection Vector
 470
 471 .. math::
 472
 473   dst.x = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.x - src1.x
 474   dst.y = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.y - src1.y
 475   dst.z = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.z - src1.z
 476   dst.w = 1.0
 477
 478 Considered for removal.
 479
 480
 481 1.5.16  SEQ - Set On Equal
 482
 483 .. math::
 484
 485   dst.x = (src0.x == src1.x) ? 1.0 : 0.0
 486   dst.y = (src0.y == src1.y) ? 1.0 : 0.0
 487   dst.z = (src0.z == src1.z) ? 1.0 : 0.0
 488   dst.w = (src0.w == src1.w) ? 1.0 : 0.0
 489
 490
 491 1.5.17  SFL - Set On False
 492
 493 .. math::
 494
 495   dst.x = 0.0
 496   dst.y = 0.0
 497   dst.z = 0.0
 498   dst.w = 0.0
 499
 500 Considered for removal.
 501
 502 1.5.18  SGT - Set On Greater Than
 503
 504 .. math::
 505
 506   dst.x = (src0.x > src1.x) ? 1.0 : 0.0
 507   dst.y = (src0.y > src1.y) ? 1.0 : 0.0
 508   dst.z = (src0.z > src1.z) ? 1.0 : 0.0
 509   dst.w = (src0.w > src1.w) ? 1.0 : 0.0
 510
 511
 512 SIN - Sine
 513
 514 .. math::
 515
 516   dst.x = \sin{src.x}
 517
 518   dst.y = \sin{src.x}
 519
 520   dst.z = \sin{src.x}
 521
 522   dst.w = \sin{src.w}
 523
 524
 525 1.5.20  SLE - Set On Less Equal Than
 526
 527 .. math::
 528
 529   dst.x = (src0.x <= src1.x) ? 1.0 : 0.0
 530   dst.y = (src0.y <= src1.y) ? 1.0 : 0.0
 531   dst.z = (src0.z <= src1.z) ? 1.0 : 0.0
 532   dst.w = (src0.w <= src1.w) ? 1.0 : 0.0
 533
 534
 535 1.5.21  SNE - Set On Not Equal
 536
 537 .. math::
 538
 539   dst.x = (src0.x != src1.x) ? 1.0 : 0.0
 540   dst.y = (src0.y != src1.y) ? 1.0 : 0.0
 541   dst.z = (src0.z != src1.z) ? 1.0 : 0.0
 542   dst.w = (src0.w != src1.w) ? 1.0 : 0.0
 543
 544
 545 1.5.22  STR - Set On True
 546
 547 .. math::
 548
 549   dst.x = 1.0
 550   dst.y = 1.0
 551   dst.z = 1.0
 552   dst.w = 1.0
 553
 554
 555 1.5.23  TEX - Texture Lookup
 556
 557   TBD
 558
 559
 560 1.5.24  TXD - Texture Lookup with Derivatives
 561
 562   TBD
 563
 564
 565 1.5.25  TXP - Projective Texture Lookup
 566
 567   TBD
 568
 569
 570 1.5.26  UP2H - Unpack Two 16-Bit Floats
 571
 572   TBD
 573
 574   Considered for removal.
 575
 576 1.5.27  UP2US - Unpack Two Unsigned 16-Bit Scalars
 577
 578   TBD
 579
 580   Considered for removal.
 581
 582 1.5.28  UP4B - Unpack Four Signed 8-Bit Values
 583
 584   TBD
 585
 586   Considered for removal.
 587
 588 1.5.29  UP4UB - Unpack Four Unsigned 8-Bit Scalars
 589
 590   TBD
 591
 592   Considered for removal.
 593
 594 1.5.30  X2D - 2D Coordinate Transformation
 595
 596 .. math::
 597
 598   dst.x = src0.x + src1.x * src2.x + src1.y * src2.y
 599   dst.y = src0.y + src1.x * src2.z + src1.y * src2.w
 600   dst.z = src0.x + src1.x * src2.x + src1.y * src2.y
 601   dst.w = src0.y + src1.x * src2.z + src1.y * src2.w
 602
 603 Considered for removal.
 604
 605
 606 1.6  GL_NV_vertex_program2
 607 --------------------------
 608
 609
 610 1.6.1  ARA - Address Register Add
 611
 612   TBD
 613
 614   Considered for removal.
 615
 616 1.6.2  ARR - Address Register Load With Round
 617
 618 .. math::
 619
 620   dst.x = round(src.x)
 621   dst.y = round(src.y)
 622   dst.z = round(src.z)
 623   dst.w = round(src.w)
 624
 625
 626 1.6.3  BRA - Branch
 627
 628   pc = target
 629
 630   Considered for removal.
 631
 632 1.6.4  CAL - Subroutine Call
 633
 634   push(pc)
 635   pc = target
 636
 637
 638 1.6.5  RET - Subroutine Call Return
 639
 640   pc = pop()
 641
 642   Potential restrictions:
 643   * Only occurs at end of function.
 644
 645 1.6.6  SSG - Set Sign
 646
 647 .. math::
 648
 649   dst.x = (src.x > 0.0) ? 1.0 : (src.x < 0.0) ? -1.0 : 0.0
 650   dst.y = (src.y > 0.0) ? 1.0 : (src.y < 0.0) ? -1.0 : 0.0
 651   dst.z = (src.z > 0.0) ? 1.0 : (src.z < 0.0) ? -1.0 : 0.0
 652   dst.w = (src.w > 0.0) ? 1.0 : (src.w < 0.0) ? -1.0 : 0.0
 653
 654
 655 1.8.1  CMP - Compare
 656
 657 .. math::
 658
 659   dst.x = (src0.x < 0.0) ? src1.x : src2.x
 660   dst.y = (src0.y < 0.0) ? src1.y : src2.y
 661   dst.z = (src0.z < 0.0) ? src1.z : src2.z
 662   dst.w = (src0.w < 0.0) ? src1.w : src2.w
 663
 664
 665 1.8.2  KIL - Conditional Discard
 666
 667 .. math::
 668
 669   if (src.x < 0.0 || src.y < 0.0 || src.z < 0.0 || src.w < 0.0)
 670     discard
 671   endif
 672
 673
 674 SCS - Sine Cosine
 675
 676 .. math::
 677
 678   dst.x = \cos{src.x}
 679
 680   dst.y = \sin{src.x}
 681
 682   dst.z = 0.0
 683
 684   dst.y = 1.0
 685
 686
 687 1.8.4  TXB - Texture Lookup With Bias
 688
 689   TBD
 690
 691
 692 1.9.1  NRM - 3-component Vector Normalise
 693
 694 .. math::
 695
 696   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z)
 697   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z)
 698   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z)
 699   dst.w = 1.0
 700
 701
 702 1.9.2  DIV - Divide
 703
 704 .. math::
 705
 706   dst.x = src0.x / src1.x
 707   dst.y = src0.y / src1.y
 708   dst.z = src0.z / src1.z
 709   dst.w = src0.w / src1.w
 710
 711
 712 1.9.3  DP2 - 2-component Dot Product
 713
 714 .. math::
 715
 716   dst.x = src0.x * src1.x + src0.y * src1.y
 717   dst.y = src0.x * src1.x + src0.y * src1.y
 718   dst.z = src0.x * src1.x + src0.y * src1.y
 719   dst.w = src0.x * src1.x + src0.y * src1.y
 720
 721
 722 1.9.5  TXL - Texture Lookup With LOD
 723
 724   TBD
 725
 726
 727 1.9.6  BRK - Break
 728
 729   TBD
 730
 731
 732 1.9.7  IF - If
 733
 734   TBD
 735
 736
 737 1.9.8  BGNFOR - Begin a For-Loop
 738
 739   dst.x = floor(src.x)
 740   dst.y = floor(src.y)
 741   dst.z = floor(src.z)
 742
 743   if (dst.y <= 0)
 744     pc = [matching ENDFOR] + 1
 745   endif
 746
 747   Note: The destination must be a loop register.
 748         The source must be a constant register.
 749
 750   Considered for cleanup / removal.
 751
 752
 753 1.9.9  REP - Repeat
 754
 755   TBD
 756
 757
 758 1.9.10  ELSE - Else
 759
 760   TBD
 761
 762
 763 1.9.11  ENDIF - End If
 764
 765   TBD
 766
 767
 768 1.9.12  ENDFOR - End a For-Loop
 769
 770   dst.x = dst.x + dst.z
 771   dst.y = dst.y - 1.0
 772
 773   if (dst.y > 0)
 774     pc = [matching BGNFOR instruction] + 1
 775   endif
 776
 777   Note: The destination must be a loop register.
 778
 779   Considered for cleanup / removal.
 780
 781 1.9.13  ENDREP - End Repeat
 782
 783   TBD
 784
 785
 786 1.10.1  PUSHA - Push Address Register On Stack
 787
 788   push(src.x)
 789   push(src.y)
 790   push(src.z)
 791   push(src.w)
 792
 793   Considered for cleanup / removal.
 794
 795 1.10.2  POPA - Pop Address Register From Stack
 796
 797   dst.w = pop()
 798   dst.z = pop()
 799   dst.y = pop()
 800   dst.x = pop()
 801
 802   Considered for cleanup / removal.
 803
 804
 805 1.11  GL_NV_gpu_program4
 806 ------------------------
 807
 808 Support for these opcodes indicated by a special pipe capability bit (TBD).
 809
 810 1.11.1  CEIL - Ceiling
 811
 812 .. math::
 813
 814   dst.x = ceil(src.x)
 815   dst.y = ceil(src.y)
 816   dst.z = ceil(src.z)
 817   dst.w = ceil(src.w)
 818
 819
 820 1.11.2  I2F - Integer To Float
 821
 822 .. math::
 823
 824   dst.x = (float) src.x
 825   dst.y = (float) src.y
 826   dst.z = (float) src.z
 827   dst.w = (float) src.w
 828
 829
 830 1.11.3  NOT - Bitwise Not
 831
 832 .. math::
 833
 834   dst.x = ~src.x
 835   dst.y = ~src.y
 836   dst.z = ~src.z
 837   dst.w = ~src.w
 838
 839
 840 1.11.4  TRUNC - Truncate
 841
 842 .. math::
 843
 844   dst.x = trunc(src.x)
 845   dst.y = trunc(src.y)
 846   dst.z = trunc(src.z)
 847   dst.w = trunc(src.w)
 848
 849
 850 1.11.5  SHL - Shift Left
 851
 852 .. math::
 853
 854   dst.x = src0.x << src1.x
 855   dst.y = src0.y << src1.x
 856   dst.z = src0.z << src1.x
 857   dst.w = src0.w << src1.x
 858
 859
 860 1.11.6  SHR - Shift Right
 861
 862 .. math::
 863
 864   dst.x = src0.x >> src1.x
 865   dst.y = src0.y >> src1.x
 866   dst.z = src0.z >> src1.x
 867   dst.w = src0.w >> src1.x
 868
 869
 870 1.11.7  AND - Bitwise And
 871
 872 .. math::
 873
 874   dst.x = src0.x & src1.x
 875   dst.y = src0.y & src1.y
 876   dst.z = src0.z & src1.z
 877   dst.w = src0.w & src1.w
 878
 879
 880 1.11.8  OR - Bitwise Or
 881
 882 .. math::
 883
 884   dst.x = src0.x | src1.x
 885   dst.y = src0.y | src1.y
 886   dst.z = src0.z | src1.z
 887   dst.w = src0.w | src1.w
 888
 889
 890 1.11.9  MOD - Modulus
 891
 892 .. math::
 893
 894   dst.x = src0.x % src1.x
 895   dst.y = src0.y % src1.y
 896   dst.z = src0.z % src1.z
 897   dst.w = src0.w % src1.w
 898
 899
 900 1.11.10  XOR - Bitwise Xor
 901
 902 .. math::
 903
 904   dst.x = src0.x ^ src1.x
 905   dst.y = src0.y ^ src1.y
 906   dst.z = src0.z ^ src1.z
 907   dst.w = src0.w ^ src1.w
 908
 909
 910 1.11.11  SAD - Sum Of Absolute Differences
 911
 912 .. math::
 913
 914   dst.x = abs(src0.x - src1.x) + src2.x
 915   dst.y = abs(src0.y - src1.y) + src2.y
 916   dst.z = abs(src0.z - src1.z) + src2.z
 917   dst.w = abs(src0.w - src1.w) + src2.w
 918
 919
 920 1.11.12  TXF - Texel Fetch
 921
 922   TBD
 923
 924
 925 1.11.13  TXQ - Texture Size Query
 926
 927   TBD
 928
 929
 930 1.11.14  CONT - Continue
 931
 932   TBD
 933
 934
 935 1.12  GL_NV_geometry_program4
 936 -----------------------------
 937
 938
 939 1.12.1  EMIT - Emit
 940
 941   TBD
 942
 943
 944 1.12.2  ENDPRIM - End Primitive
 945
 946   TBD
 947
 948
 949 1.13  GLSL
 950 ----------
 951
 952
 953 1.13.1  BGNLOOP - Begin a Loop
 954
 955   TBD
 956
 957
 958 1.13.2  BGNSUB - Begin Subroutine
 959
 960   TBD
 961
 962
 963 1.13.3  ENDLOOP - End a Loop
 964
 965   TBD
 966
 967
 968 1.13.4  ENDSUB - End Subroutine
 969
 970   TBD
 971
 972
 973
 974 1.13.10  NOP - No Operation
 975
 976   Do nothing.
 977
 978
 979
 980 1.16.7  NRM4 - 4-component Vector Normalise
 981
 982 .. math::
 983
 984   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 985   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 986   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 987   dst.w = src.w / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 988
 989
 990 1.17  ps_2_x
 991 ------------
 992
 993
 994 1.17.2  CALLNZ - Subroutine Call If Not Zero
 995
 996   TBD
 997
 998
 999 1.17.3  IFC - If
1000
1001   TBD
1002
1003
1004 1.17.5  BREAKC - Break Conditional
1005
1006   TBD
1007
1008
1009 2  Explanation of symbols used
1010 ==============================
1011
1012
1013 2.1  Functions
1014 --------------
1015
1016
1017   abs(x)            Absolute value of x.
1018                     (x < 0.0) ? -x : x
1019
1020   ceil(x)           Ceiling of x.
1021
1022   clamp(x,y,z)      Clamp x between y and z.
1023                     (x < y) ? y : (x > z) ? z : x
1024
1025   :math:`\lfloor x\rfloor` Floor of `x`.
1026
1027   lg2(x)            Logarithm base 2 of x.
1028
1029   max(x,y)          Maximum of x and y.
1030                     (x > y) ? x : y
1031
1032   min(x,y)          Minimum of x and y.
1033                     (x < y) ? x : y
1034
1035   partialx(x)       Derivative of x relative to fragment's X.
1036
1037   partialy(x)       Derivative of x relative to fragment's Y.
1038
1039   pop()             Pop from stack.
1040
1041   :math:`x^y`       `x` to the power `y`.
1042
1043   push(x)           Push x on stack.
1044
1045   round(x)          Round x.
1046
1047   trunc(x)          Truncate x.
1048
1049
1050 2.2  Keywords
1051 -------------
1052
1053
1054   discard           Discard fragment.
1055
1056   dst               First destination register.
1057
1058   dst0              First destination register.
1059
1060   pc                Program counter.
1061
1062   src               First source register.
1063
1064   src0              First source register.
1065
1066   src1              Second source register.
1067
1068   src2              Third source register.
1069
1070   target            Label of target instruction.
1071
1072
1073 3  Other tokens
1074 ===============
1075
1076
1077 3.1  Declaration Semantic
1078 -------------------------
1079
1080
1081   Follows Declaration token if Semantic bit is set.
1082
1083   Since its purpose is to link a shader with other stages of the pipeline,
1084   it is valid to follow only those Declaration tokens that declare a register
1085   either in INPUT or OUTPUT file.
1086
1087   SemanticName field contains the semantic name of the register being declared.
1088   There is no default value.
1089
1090   SemanticIndex is an optional subscript that can be used to distinguish
1091   different register declarations with the same semantic name. The default value
1092   is 0.
1093
1094   The meanings of the individual semantic names are explained in the following
1095   sections.
1096
1097
1098 3.1.1  FACE
1099
1100   Valid only in a fragment shader INPUT declaration.
1101
1102   FACE.x is negative when the primitive is back facing. FACE.x is positive
1103   when the primitive is front facing.