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 = floor(src.x)
  18
  19   dst.y = floor(src.y)
  20
  21   dst.z = floor(src.z)
  22
  23   dst.w = floor(src.w)
  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) ? pow(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 = pow(2.0, floor(src.x))
  83
  84   dst.y = src.x - floor(src.x)
  85
  86   dst.z = pow(2.0, src.x)
  87
  88   dst.w = 1.0
  89
  90
  91 LOG - Approximate Logarithm Base 2
  92
  93 .. math::
  94
  95   dst.x = floor(lg2(abs(src.x)))
  96
  97   dst.y = abs(src.x) / pow(2.0, floor(lg2(abs(src.x))))
  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 - floor(src.x)
 291
 292   dst.y = src.y - floor(src.y)
 293
 294   dst.z = src.z - floor(src.z)
 295
 296   dst.w = src.w - floor(src.w)
 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 1.3.8  FLR - Floor
 310
 311 .. math::
 312
 313   dst.x = floor(src.x)
 314   dst.y = floor(src.y)
 315   dst.z = floor(src.z)
 316   dst.w = floor(src.w)
 317
 318
 319 1.3.9  ROUND - Round
 320
 321 .. math::
 322
 323   dst.x = round(src.x)
 324   dst.y = round(src.y)
 325   dst.z = round(src.z)
 326   dst.w = round(src.w)
 327
 328
 329 1.3.10  EX2 - Exponential Base 2
 330
 331 .. math::
 332
 333   dst.x = pow(2.0, src.x)
 334   dst.y = pow(2.0, src.x)
 335   dst.z = pow(2.0, src.x)
 336   dst.w = pow(2.0, src.x)
 337
 338
 339 1.3.11  LG2 - Logarithm Base 2
 340
 341 .. math::
 342
 343   dst.x = lg2(src.x)
 344   dst.y = lg2(src.x)
 345   dst.z = lg2(src.x)
 346   dst.w = lg2(src.x)
 347
 348
 349 1.3.12  POW - Power
 350
 351 .. math::
 352
 353   dst.x = pow(src0.x, src1.x)
 354   dst.y = pow(src0.x, src1.x)
 355   dst.z = pow(src0.x, src1.x)
 356   dst.w = pow(src0.x, src1.x)
 357
 358 1.3.15  XPD - Cross Product
 359
 360 .. math::
 361
 362   dst.x = src0.y * src1.z - src1.y * src0.z
 363   dst.y = src0.z * src1.x - src1.z * src0.x
 364   dst.z = src0.x * src1.y - src1.x * src0.y
 365   dst.w = 1.0
 366
 367
 368 1.4.1  ABS - Absolute
 369
 370 .. math::
 371
 372   dst.x = abs(src.x)
 373   dst.y = abs(src.y)
 374   dst.z = abs(src.z)
 375   dst.w = abs(src.w)
 376
 377
 378 1.4.2  RCC - Reciprocal Clamped
 379
 380 .. math::
 381
 382   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)
 383   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)
 384   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)
 385   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)
 386
 387
 388 1.4.3  DPH - Homogeneous Dot Product
 389
 390 .. math::
 391
 392   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 393   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 394   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 395   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 396
 397
 398 1.5.1  COS - Cosine
 399
 400 .. math::
 401
 402   dst.x = cos(src.x)
 403   dst.y = cos(src.x)
 404   dst.z = cos(src.x)
 405   dst.w = cos(src.w)
 406
 407
 408 1.5.2  DDX - Derivative Relative To X
 409
 410 .. math::
 411
 412   dst.x = partialx(src.x)
 413   dst.y = partialx(src.y)
 414   dst.z = partialx(src.z)
 415   dst.w = partialx(src.w)
 416
 417
 418 1.5.3  DDY - Derivative Relative To Y
 419
 420 .. math::
 421
 422   dst.x = partialy(src.x)
 423   dst.y = partialy(src.y)
 424   dst.z = partialy(src.z)
 425   dst.w = partialy(src.w)
 426
 427
 428 1.5.7  KILP - Predicated Discard
 429
 430 .. math::
 431
 432   discard
 433
 434
 435 1.5.10  PK2H - Pack Two 16-bit Floats
 436
 437   TBD
 438
 439
 440 1.5.11  PK2US - Pack Two Unsigned 16-bit Scalars
 441
 442   TBD
 443
 444
 445 1.5.12  PK4B - Pack Four Signed 8-bit Scalars
 446
 447   TBD
 448
 449
 450 1.5.13  PK4UB - Pack Four Unsigned 8-bit Scalars
 451
 452   TBD
 453
 454
 455 1.5.15  RFL - Reflection Vector
 456
 457 .. math::
 458
 459   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
 460   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
 461   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
 462   dst.w = 1.0
 463
 464 Considered for removal.
 465
 466
 467 1.5.16  SEQ - Set On Equal
 468
 469 .. math::
 470
 471   dst.x = (src0.x == src1.x) ? 1.0 : 0.0
 472   dst.y = (src0.y == src1.y) ? 1.0 : 0.0
 473   dst.z = (src0.z == src1.z) ? 1.0 : 0.0
 474   dst.w = (src0.w == src1.w) ? 1.0 : 0.0
 475
 476
 477 1.5.17  SFL - Set On False
 478
 479 .. math::
 480
 481   dst.x = 0.0
 482   dst.y = 0.0
 483   dst.z = 0.0
 484   dst.w = 0.0
 485
 486 Considered for removal.
 487
 488 1.5.18  SGT - Set On Greater Than
 489
 490 .. math::
 491
 492   dst.x = (src0.x > src1.x) ? 1.0 : 0.0
 493   dst.y = (src0.y > src1.y) ? 1.0 : 0.0
 494   dst.z = (src0.z > src1.z) ? 1.0 : 0.0
 495   dst.w = (src0.w > src1.w) ? 1.0 : 0.0
 496
 497
 498 1.5.19  SIN - Sine
 499
 500 .. math::
 501
 502   dst.x = sin(src.x)
 503   dst.y = sin(src.x)
 504   dst.z = sin(src.x)
 505   dst.w = sin(src.w)
 506
 507
 508 1.5.20  SLE - Set On Less Equal Than
 509
 510 .. math::
 511
 512   dst.x = (src0.x <= src1.x) ? 1.0 : 0.0
 513   dst.y = (src0.y <= src1.y) ? 1.0 : 0.0
 514   dst.z = (src0.z <= src1.z) ? 1.0 : 0.0
 515   dst.w = (src0.w <= src1.w) ? 1.0 : 0.0
 516
 517
 518 1.5.21  SNE - Set On Not Equal
 519
 520 .. math::
 521
 522   dst.x = (src0.x != src1.x) ? 1.0 : 0.0
 523   dst.y = (src0.y != src1.y) ? 1.0 : 0.0
 524   dst.z = (src0.z != src1.z) ? 1.0 : 0.0
 525   dst.w = (src0.w != src1.w) ? 1.0 : 0.0
 526
 527
 528 1.5.22  STR - Set On True
 529
 530 .. math::
 531
 532   dst.x = 1.0
 533   dst.y = 1.0
 534   dst.z = 1.0
 535   dst.w = 1.0
 536
 537
 538 1.5.23  TEX - Texture Lookup
 539
 540   TBD
 541
 542
 543 1.5.24  TXD - Texture Lookup with Derivatives
 544
 545   TBD
 546
 547
 548 1.5.25  TXP - Projective Texture Lookup
 549
 550   TBD
 551
 552
 553 1.5.26  UP2H - Unpack Two 16-Bit Floats
 554
 555   TBD
 556
 557   Considered for removal.
 558
 559 1.5.27  UP2US - Unpack Two Unsigned 16-Bit Scalars
 560
 561   TBD
 562
 563   Considered for removal.
 564
 565 1.5.28  UP4B - Unpack Four Signed 8-Bit Values
 566
 567   TBD
 568
 569   Considered for removal.
 570
 571 1.5.29  UP4UB - Unpack Four Unsigned 8-Bit Scalars
 572
 573   TBD
 574
 575   Considered for removal.
 576
 577 1.5.30  X2D - 2D Coordinate Transformation
 578
 579 .. math::
 580
 581   dst.x = src0.x + src1.x * src2.x + src1.y * src2.y
 582   dst.y = src0.y + src1.x * src2.z + src1.y * src2.w
 583   dst.z = src0.x + src1.x * src2.x + src1.y * src2.y
 584   dst.w = src0.y + src1.x * src2.z + src1.y * src2.w
 585
 586 Considered for removal.
 587
 588
 589 1.6  GL_NV_vertex_program2
 590 --------------------------
 591
 592
 593 1.6.1  ARA - Address Register Add
 594
 595   TBD
 596
 597   Considered for removal.
 598
 599 1.6.2  ARR - Address Register Load With Round
 600
 601 .. math::
 602
 603   dst.x = round(src.x)
 604   dst.y = round(src.y)
 605   dst.z = round(src.z)
 606   dst.w = round(src.w)
 607
 608
 609 1.6.3  BRA - Branch
 610
 611   pc = target
 612
 613   Considered for removal.
 614
 615 1.6.4  CAL - Subroutine Call
 616
 617   push(pc)
 618   pc = target
 619
 620
 621 1.6.5  RET - Subroutine Call Return
 622
 623   pc = pop()
 624
 625   Potential restrictions:
 626   * Only occurs at end of function.
 627
 628 1.6.6  SSG - Set Sign
 629
 630 .. math::
 631
 632   dst.x = (src.x > 0.0) ? 1.0 : (src.x < 0.0) ? -1.0 : 0.0
 633   dst.y = (src.y > 0.0) ? 1.0 : (src.y < 0.0) ? -1.0 : 0.0
 634   dst.z = (src.z > 0.0) ? 1.0 : (src.z < 0.0) ? -1.0 : 0.0
 635   dst.w = (src.w > 0.0) ? 1.0 : (src.w < 0.0) ? -1.0 : 0.0
 636
 637
 638 1.8.1  CMP - Compare
 639
 640 .. math::
 641
 642   dst.x = (src0.x < 0.0) ? src1.x : src2.x
 643   dst.y = (src0.y < 0.0) ? src1.y : src2.y
 644   dst.z = (src0.z < 0.0) ? src1.z : src2.z
 645   dst.w = (src0.w < 0.0) ? src1.w : src2.w
 646
 647
 648 1.8.2  KIL - Conditional Discard
 649
 650 .. math::
 651
 652   if (src.x < 0.0 || src.y < 0.0 || src.z < 0.0 || src.w < 0.0)
 653     discard
 654   endif
 655
 656
 657 1.8.3  SCS - Sine Cosine
 658
 659 .. math::
 660
 661   dst.x = cos(src.x)
 662   dst.y = sin(src.x)
 663   dst.z = 0.0
 664   dst.y = 1.0
 665
 666
 667 1.8.4  TXB - Texture Lookup With Bias
 668
 669   TBD
 670
 671
 672 1.9.1  NRM - 3-component Vector Normalise
 673
 674 .. math::
 675
 676   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z)
 677   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z)
 678   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z)
 679   dst.w = 1.0
 680
 681
 682 1.9.2  DIV - Divide
 683
 684 .. math::
 685
 686   dst.x = src0.x / src1.x
 687   dst.y = src0.y / src1.y
 688   dst.z = src0.z / src1.z
 689   dst.w = src0.w / src1.w
 690
 691
 692 1.9.3  DP2 - 2-component Dot Product
 693
 694 .. math::
 695
 696   dst.x = src0.x * src1.x + src0.y * src1.y
 697   dst.y = src0.x * src1.x + src0.y * src1.y
 698   dst.z = src0.x * src1.x + src0.y * src1.y
 699   dst.w = src0.x * src1.x + src0.y * src1.y
 700
 701
 702 1.9.5  TXL - Texture Lookup With LOD
 703
 704   TBD
 705
 706
 707 1.9.6  BRK - Break
 708
 709   TBD
 710
 711
 712 1.9.7  IF - If
 713
 714   TBD
 715
 716
 717 1.9.8  BGNFOR - Begin a For-Loop
 718
 719   dst.x = floor(src.x)
 720   dst.y = floor(src.y)
 721   dst.z = floor(src.z)
 722
 723   if (dst.y <= 0)
 724     pc = [matching ENDFOR] + 1
 725   endif
 726
 727   Note: The destination must be a loop register.
 728         The source must be a constant register.
 729
 730   Considered for cleanup / removal.
 731
 732
 733 1.9.9  REP - Repeat
 734
 735   TBD
 736
 737
 738 1.9.10  ELSE - Else
 739
 740   TBD
 741
 742
 743 1.9.11  ENDIF - End If
 744
 745   TBD
 746
 747
 748 1.9.12  ENDFOR - End a For-Loop
 749
 750   dst.x = dst.x + dst.z
 751   dst.y = dst.y - 1.0
 752
 753   if (dst.y > 0)
 754     pc = [matching BGNFOR instruction] + 1
 755   endif
 756
 757   Note: The destination must be a loop register.
 758
 759   Considered for cleanup / removal.
 760
 761 1.9.13  ENDREP - End Repeat
 762
 763   TBD
 764
 765
 766 1.10.1  PUSHA - Push Address Register On Stack
 767
 768   push(src.x)
 769   push(src.y)
 770   push(src.z)
 771   push(src.w)
 772
 773   Considered for cleanup / removal.
 774
 775 1.10.2  POPA - Pop Address Register From Stack
 776
 777   dst.w = pop()
 778   dst.z = pop()
 779   dst.y = pop()
 780   dst.x = pop()
 781
 782   Considered for cleanup / removal.
 783
 784
 785 1.11  GL_NV_gpu_program4
 786 ------------------------
 787
 788 Support for these opcodes indicated by a special pipe capability bit (TBD).
 789
 790 1.11.1  CEIL - Ceiling
 791
 792 .. math::
 793
 794   dst.x = ceil(src.x)
 795   dst.y = ceil(src.y)
 796   dst.z = ceil(src.z)
 797   dst.w = ceil(src.w)
 798
 799
 800 1.11.2  I2F - Integer To Float
 801
 802 .. math::
 803
 804   dst.x = (float) src.x
 805   dst.y = (float) src.y
 806   dst.z = (float) src.z
 807   dst.w = (float) src.w
 808
 809
 810 1.11.3  NOT - Bitwise Not
 811
 812 .. math::
 813
 814   dst.x = ~src.x
 815   dst.y = ~src.y
 816   dst.z = ~src.z
 817   dst.w = ~src.w
 818
 819
 820 1.11.4  TRUNC - Truncate
 821
 822 .. math::
 823
 824   dst.x = trunc(src.x)
 825   dst.y = trunc(src.y)
 826   dst.z = trunc(src.z)
 827   dst.w = trunc(src.w)
 828
 829
 830 1.11.5  SHL - Shift Left
 831
 832 .. math::
 833
 834   dst.x = src0.x << src1.x
 835   dst.y = src0.y << src1.x
 836   dst.z = src0.z << src1.x
 837   dst.w = src0.w << src1.x
 838
 839
 840 1.11.6  SHR - Shift Right
 841
 842 .. math::
 843
 844   dst.x = src0.x >> src1.x
 845   dst.y = src0.y >> src1.x
 846   dst.z = src0.z >> src1.x
 847   dst.w = src0.w >> src1.x
 848
 849
 850 1.11.7  AND - Bitwise And
 851
 852 .. math::
 853
 854   dst.x = src0.x & src1.x
 855   dst.y = src0.y & src1.y
 856   dst.z = src0.z & src1.z
 857   dst.w = src0.w & src1.w
 858
 859
 860 1.11.8  OR - Bitwise Or
 861
 862 .. math::
 863
 864   dst.x = src0.x | src1.x
 865   dst.y = src0.y | src1.y
 866   dst.z = src0.z | src1.z
 867   dst.w = src0.w | src1.w
 868
 869
 870 1.11.9  MOD - Modulus
 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 1.11.10  XOR - Bitwise Xor
 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 1.11.11  SAD - Sum Of Absolute Differences
 891
 892 .. math::
 893
 894   dst.x = abs(src0.x - src1.x) + src2.x
 895   dst.y = abs(src0.y - src1.y) + src2.y
 896   dst.z = abs(src0.z - src1.z) + src2.z
 897   dst.w = abs(src0.w - src1.w) + src2.w
 898
 899
 900 1.11.12  TXF - Texel Fetch
 901
 902   TBD
 903
 904
 905 1.11.13  TXQ - Texture Size Query
 906
 907   TBD
 908
 909
 910 1.11.14  CONT - Continue
 911
 912   TBD
 913
 914
 915 1.12  GL_NV_geometry_program4
 916 -----------------------------
 917
 918
 919 1.12.1  EMIT - Emit
 920
 921   TBD
 922
 923
 924 1.12.2  ENDPRIM - End Primitive
 925
 926   TBD
 927
 928
 929 1.13  GLSL
 930 ----------
 931
 932
 933 1.13.1  BGNLOOP - Begin a Loop
 934
 935   TBD
 936
 937
 938 1.13.2  BGNSUB - Begin Subroutine
 939
 940   TBD
 941
 942
 943 1.13.3  ENDLOOP - End a Loop
 944
 945   TBD
 946
 947
 948 1.13.4  ENDSUB - End Subroutine
 949
 950   TBD
 951
 952
 953
 954 1.13.10  NOP - No Operation
 955
 956   Do nothing.
 957
 958
 959
 960 1.16.7  NRM4 - 4-component Vector Normalise
 961
 962 .. math::
 963
 964   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 965   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 966   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 967   dst.w = src.w / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 968
 969
 970 1.17  ps_2_x
 971 ------------
 972
 973
 974 1.17.2  CALLNZ - Subroutine Call If Not Zero
 975
 976   TBD
 977
 978
 979 1.17.3  IFC - If
 980
 981   TBD
 982
 983
 984 1.17.5  BREAKC - Break Conditional
 985
 986   TBD
 987
 988
 989 2  Explanation of symbols used
 990 ==============================
 991
 992
 993 2.1  Functions
 994 --------------
 995
 996
 997   abs(x)            Absolute value of x.
 998                     (x < 0.0) ? -x : x
 999
1000   ceil(x)           Ceiling of x.
1001
1002   clamp(x,y,z)      Clamp x between y and z.
1003                     (x < y) ? y : (x > z) ? z : x
1004
1005   cos(x)            Cosine of x.
1006
1007   floor(x)          Floor of x.
1008
1009   lg2(x)            Logarithm base 2 of x.
1010
1011   max(x,y)          Maximum of x and y.
1012                     (x > y) ? x : y
1013
1014   min(x,y)          Minimum of x and y.
1015                     (x < y) ? x : y
1016
1017   partialx(x)       Derivative of x relative to fragment's X.
1018
1019   partialy(x)       Derivative of x relative to fragment's Y.
1020
1021   pop()             Pop from stack.
1022
1023   pow(x,y)          Raise x to power of y.
1024
1025   push(x)           Push x on stack.
1026
1027   round(x)          Round x.
1028
1029   sin(x)            Sine of x.
1030
1031   sqrt(x)           Square root of x.
1032
1033   trunc(x)          Truncate x.
1034
1035
1036 2.2  Keywords
1037 -------------
1038
1039
1040   discard           Discard fragment.
1041
1042   dst               First destination register.
1043
1044   dst0              First destination register.
1045
1046   pc                Program counter.
1047
1048   src               First source register.
1049
1050   src0              First source register.
1051
1052   src1              Second source register.
1053
1054   src2              Third source register.
1055
1056   target            Label of target instruction.
1057
1058
1059 3  Other tokens
1060 ===============
1061
1062
1063 3.1  Declaration Semantic
1064 -------------------------
1065
1066
1067   Follows Declaration token if Semantic bit is set.
1068
1069   Since its purpose is to link a shader with other stages of the pipeline,
1070   it is valid to follow only those Declaration tokens that declare a register
1071   either in INPUT or OUTPUT file.
1072
1073   SemanticName field contains the semantic name of the register being declared.
1074   There is no default value.
1075
1076   SemanticIndex is an optional subscript that can be used to distinguish
1077   different register declarations with the same semantic name. The default value
1078   is 0.
1079
1080   The meanings of the individual semantic names are explained in the following
1081   sections.
1082
1083
1084 3.1.1  FACE
1085
1086   Valid only in a fragment shader INPUT declaration.
1087
1088   FACE.x is negative when the primitive is back facing. FACE.x is positive
1089   when the primitive is front facing.