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
  10 TGSI Instruction Specification
  11 ==============================
  12
  13
  14 1  Instruction Set Operations
  15 =============================
  16
  17
  18 1.1  GL_NV_vertex_program
  19 -------------------------
  20
  21
  22 1.1.1  ARL - Address Register Load
  23
  24   dst.x = floor(src.x)
  25   dst.y = floor(src.y)
  26   dst.z = floor(src.z)
  27   dst.w = floor(src.w)
  28
  29
  30 1.1.2  MOV - Move
  31
  32   dst.x = src.x
  33   dst.y = src.y
  34   dst.z = src.z
  35   dst.w = src.w
  36
  37
  38 1.1.3  LIT - Light Coefficients
  39
  40   dst.x = 1.0
  41   dst.y = max(src.x, 0.0)
  42   dst.z = (src.x > 0.0) ? pow(max(src.y, 0.0), clamp(src.w, -128.0, 128.0)) : 0.0
  43   dst.w = 1.0
  44
  45
  46 1.1.4  RCP - Reciprocal
  47
  48   dst.x = 1.0 / src.x
  49   dst.y = 1.0 / src.x
  50   dst.z = 1.0 / src.x
  51   dst.w = 1.0 / src.x
  52
  53
  54 1.1.5  RSQ - Reciprocal Square Root
  55
  56   dst.x = 1.0 / sqrt(abs(src.x))
  57   dst.y = 1.0 / sqrt(abs(src.x))
  58   dst.z = 1.0 / sqrt(abs(src.x))
  59   dst.w = 1.0 / sqrt(abs(src.x))
  60
  61
  62 1.1.6  EXP - Approximate Exponential Base 2
  63
  64   dst.x = pow(2.0, floor(src.x))
  65   dst.y = src.x - floor(src.x)
  66   dst.z = pow(2.0, src.x)
  67   dst.w = 1.0
  68
  69
  70 1.1.7  LOG - Approximate Logarithm Base 2
  71
  72   dst.x = floor(lg2(abs(src.x)))
  73   dst.y = abs(src.x) / pow(2.0, floor(lg2(abs(src.x))))
  74   dst.z = lg2(abs(src.x))
  75   dst.w = 1.0
  76
  77
  78 1.1.8  MUL - Multiply
  79
  80   dst.x = src0.x * src1.x
  81   dst.y = src0.y * src1.y
  82   dst.z = src0.z * src1.z
  83   dst.w = src0.w * src1.w
  84
  85
  86 1.1.9  ADD - Add
  87
  88   dst.x = src0.x + src1.x
  89   dst.y = src0.y + src1.y
  90   dst.z = src0.z + src1.z
  91   dst.w = src0.w + src1.w
  92
  93
  94 1.1.10  DP3 - 3-component Dot Product
  95
  96   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
  97   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
  98   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
  99   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
 100
 101
 102 1.1.11  DP4 - 4-component Dot Product
 103
 104   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 105   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 106   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 107   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
 108
 109
 110 1.1.12  DST - Distance Vector
 111
 112   dst.x = 1.0
 113   dst.y = src0.y * src1.y
 114   dst.z = src0.z
 115   dst.w = src1.w
 116
 117
 118 1.1.13  MIN - Minimum
 119
 120   dst.x = min(src0.x, src1.x)
 121   dst.y = min(src0.y, src1.y)
 122   dst.z = min(src0.z, src1.z)
 123   dst.w = min(src0.w, src1.w)
 124
 125
 126 1.1.14  MAX - Maximum
 127
 128   dst.x = max(src0.x, src1.x)
 129   dst.y = max(src0.y, src1.y)
 130   dst.z = max(src0.z, src1.z)
 131   dst.w = max(src0.w, src1.w)
 132
 133
 134 1.1.15  SLT - Set On Less Than
 135
 136   dst.x = (src0.x < src1.x) ? 1.0 : 0.0
 137   dst.y = (src0.y < src1.y) ? 1.0 : 0.0
 138   dst.z = (src0.z < src1.z) ? 1.0 : 0.0
 139   dst.w = (src0.w < src1.w) ? 1.0 : 0.0
 140
 141
 142 1.1.16  SGE - Set On Greater Equal Than
 143
 144   dst.x = (src0.x >= src1.x) ? 1.0 : 0.0
 145   dst.y = (src0.y >= src1.y) ? 1.0 : 0.0
 146   dst.z = (src0.z >= src1.z) ? 1.0 : 0.0
 147   dst.w = (src0.w >= src1.w) ? 1.0 : 0.0
 148
 149
 150 1.1.17  MAD - Multiply And Add
 151
 152   dst.x = src0.x * src1.x + src2.x
 153   dst.y = src0.y * src1.y + src2.y
 154   dst.z = src0.z * src1.z + src2.z
 155   dst.w = src0.w * src1.w + src2.w
 156
 157
 158 1.2.1  SUB - Subtract
 159
 160   dst.x = src0.x - src1.x
 161   dst.y = src0.y - src1.y
 162   dst.z = src0.z - src1.z
 163   dst.w = src0.w - src1.w
 164
 165
 166 1.2.4  LRP - Linear Interpolate
 167
 168   dst.x = src0.x * (src1.x - src2.x) + src2.x
 169   dst.y = src0.y * (src1.y - src2.y) + src2.y
 170   dst.z = src0.z * (src1.z - src2.z) + src2.z
 171   dst.w = src0.w * (src1.w - src2.w) + src2.w
 172
 173
 174 1.2.5  CND - Condition
 175
 176   dst.x = (src2.x > 0.5) ? src0.x : src1.x
 177   dst.y = (src2.y > 0.5) ? src0.y : src1.y
 178   dst.z = (src2.z > 0.5) ? src0.z : src1.z
 179   dst.w = (src2.w > 0.5) ? src0.w : src1.w
 180
 181
 182 1.2.7  DP2A - 2-component Dot Product And Add
 183
 184   dst.x = src0.x * src1.x + src0.y * src1.y + src2.x
 185   dst.y = src0.x * src1.x + src0.y * src1.y + src2.x
 186   dst.z = src0.x * src1.x + src0.y * src1.y + src2.x
 187   dst.w = src0.x * src1.x + src0.y * src1.y + src2.x
 188
 189
 190 1.3.4  FRAC - Fraction
 191
 192   dst.x = src.x - floor(src.x)
 193   dst.y = src.y - floor(src.y)
 194   dst.z = src.z - floor(src.z)
 195   dst.w = src.w - floor(src.w)
 196
 197
 198 1.3.7  CLAMP - Clamp
 199
 200   dst.x = clamp(src0.x, src1.x, src2.x)
 201   dst.y = clamp(src0.y, src1.y, src2.y)
 202   dst.z = clamp(src0.z, src1.z, src2.z)
 203   dst.w = clamp(src0.w, src1.w, src2.w)
 204
 205
 206 1.3.8  FLR - Floor
 207
 208   dst.x = floor(src.x)
 209   dst.y = floor(src.y)
 210   dst.z = floor(src.z)
 211   dst.w = floor(src.w)
 212
 213
 214 1.3.9  ROUND - Round
 215
 216   dst.x = round(src.x)
 217   dst.y = round(src.y)
 218   dst.z = round(src.z)
 219   dst.w = round(src.w)
 220
 221
 222 1.3.10  EX2 - Exponential Base 2
 223
 224   dst.x = pow(2.0, src.x)
 225   dst.y = pow(2.0, src.x)
 226   dst.z = pow(2.0, src.x)
 227   dst.w = pow(2.0, src.x)
 228
 229
 230 1.3.11  LG2 - Logarithm Base 2
 231
 232   dst.x = lg2(src.x)
 233   dst.y = lg2(src.x)
 234   dst.z = lg2(src.x)
 235   dst.w = lg2(src.x)
 236
 237
 238 1.3.12  POW - Power
 239
 240   dst.x = pow(src0.x, src1.x)
 241   dst.y = pow(src0.x, src1.x)
 242   dst.z = pow(src0.x, src1.x)
 243   dst.w = pow(src0.x, src1.x)
 244
 245 1.3.15  XPD - Cross Product
 246
 247   dst.x = src0.y * src1.z - src1.y * src0.z
 248   dst.y = src0.z * src1.x - src1.z * src0.x
 249   dst.z = src0.x * src1.y - src1.x * src0.y
 250   dst.w = 1.0
 251
 252
 253 1.4.1  ABS - Absolute
 254
 255   dst.x = abs(src.x)
 256   dst.y = abs(src.y)
 257   dst.z = abs(src.z)
 258   dst.w = abs(src.w)
 259
 260
 261 1.4.2  RCC - Reciprocal Clamped
 262
 263   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)
 264   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)
 265   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)
 266   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)
 267
 268
 269 1.4.3  DPH - Homogeneous Dot Product
 270
 271   dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 272   dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 273   dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 274   dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
 275
 276
 277 1.5.1  COS - Cosine
 278
 279   dst.x = cos(src.x)
 280   dst.y = cos(src.x)
 281   dst.z = cos(src.x)
 282   dst.w = cos(src.w)
 283
 284
 285 1.5.2  DDX - Derivative Relative To X
 286
 287   dst.x = partialx(src.x)
 288   dst.y = partialx(src.y)
 289   dst.z = partialx(src.z)
 290   dst.w = partialx(src.w)
 291
 292
 293 1.5.3  DDY - Derivative Relative To Y
 294
 295   dst.x = partialy(src.x)
 296   dst.y = partialy(src.y)
 297   dst.z = partialy(src.z)
 298   dst.w = partialy(src.w)
 299
 300
 301 1.5.7  KILP - Predicated Discard
 302
 303   discard
 304
 305
 306 1.5.10  PK2H - Pack Two 16-bit Floats
 307
 308   TBD
 309
 310
 311 1.5.11  PK2US - Pack Two Unsigned 16-bit Scalars
 312
 313   TBD
 314
 315
 316 1.5.12  PK4B - Pack Four Signed 8-bit Scalars
 317
 318   TBD
 319
 320
 321 1.5.13  PK4UB - Pack Four Unsigned 8-bit Scalars
 322
 323   TBD
 324
 325
 326 1.5.15  RFL - Reflection Vector
 327
 328   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
 329   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
 330   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
 331   dst.w = 1.0
 332
 333  Considered for removal.
 334
 335
 336 1.5.16  SEQ - Set On Equal
 337
 338   dst.x = (src0.x == src1.x) ? 1.0 : 0.0
 339   dst.y = (src0.y == src1.y) ? 1.0 : 0.0
 340   dst.z = (src0.z == src1.z) ? 1.0 : 0.0
 341   dst.w = (src0.w == src1.w) ? 1.0 : 0.0
 342
 343
 344 1.5.17  SFL - Set On False
 345
 346   dst.x = 0.0
 347   dst.y = 0.0
 348   dst.z = 0.0
 349   dst.w = 0.0
 350
 351  Considered for removal.
 352
 353 1.5.18  SGT - Set On Greater Than
 354
 355   dst.x = (src0.x > src1.x) ? 1.0 : 0.0
 356   dst.y = (src0.y > src1.y) ? 1.0 : 0.0
 357   dst.z = (src0.z > src1.z) ? 1.0 : 0.0
 358   dst.w = (src0.w > src1.w) ? 1.0 : 0.0
 359
 360
 361 1.5.19  SIN - Sine
 362
 363   dst.x = sin(src.x)
 364   dst.y = sin(src.x)
 365   dst.z = sin(src.x)
 366   dst.w = sin(src.w)
 367
 368
 369 1.5.20  SLE - Set On Less Equal Than
 370
 371   dst.x = (src0.x <= src1.x) ? 1.0 : 0.0
 372   dst.y = (src0.y <= src1.y) ? 1.0 : 0.0
 373   dst.z = (src0.z <= src1.z) ? 1.0 : 0.0
 374   dst.w = (src0.w <= src1.w) ? 1.0 : 0.0
 375
 376
 377 1.5.21  SNE - Set On Not Equal
 378
 379   dst.x = (src0.x != src1.x) ? 1.0 : 0.0
 380   dst.y = (src0.y != src1.y) ? 1.0 : 0.0
 381   dst.z = (src0.z != src1.z) ? 1.0 : 0.0
 382   dst.w = (src0.w != src1.w) ? 1.0 : 0.0
 383
 384
 385 1.5.22  STR - Set On True
 386
 387   dst.x = 1.0
 388   dst.y = 1.0
 389   dst.z = 1.0
 390   dst.w = 1.0
 391
 392
 393 1.5.23  TEX - Texture Lookup
 394
 395   TBD
 396
 397
 398 1.5.24  TXD - Texture Lookup with Derivatives
 399
 400   TBD
 401
 402
 403 1.5.25  TXP - Projective Texture Lookup
 404
 405   TBD
 406
 407
 408 1.5.26  UP2H - Unpack Two 16-Bit Floats
 409
 410   TBD
 411
 412   Considered for removal.
 413
 414 1.5.27  UP2US - Unpack Two Unsigned 16-Bit Scalars
 415
 416   TBD
 417
 418   Considered for removal.
 419
 420 1.5.28  UP4B - Unpack Four Signed 8-Bit Values
 421
 422   TBD
 423
 424   Considered for removal.
 425
 426 1.5.29  UP4UB - Unpack Four Unsigned 8-Bit Scalars
 427
 428   TBD
 429
 430   Considered for removal.
 431
 432 1.5.30  X2D - 2D Coordinate Transformation
 433
 434   dst.x = src0.x + src1.x * src2.x + src1.y * src2.y
 435   dst.y = src0.y + src1.x * src2.z + src1.y * src2.w
 436   dst.z = src0.x + src1.x * src2.x + src1.y * src2.y
 437   dst.w = src0.y + src1.x * src2.z + src1.y * src2.w
 438
 439   Considered for removal.
 440
 441
 442 1.6  GL_NV_vertex_program2
 443 --------------------------
 444
 445
 446 1.6.1  ARA - Address Register Add
 447
 448   TBD
 449
 450   Considered for removal.
 451
 452 1.6.2  ARR - Address Register Load With Round
 453
 454   dst.x = round(src.x)
 455   dst.y = round(src.y)
 456   dst.z = round(src.z)
 457   dst.w = round(src.w)
 458
 459
 460 1.6.3  BRA - Branch
 461
 462   pc = target
 463
 464   Considered for removal.
 465
 466 1.6.4  CAL - Subroutine Call
 467
 468   push(pc)
 469   pc = target
 470
 471
 472 1.6.5  RET - Subroutine Call Return
 473
 474   pc = pop()
 475
 476   Potential restrictions:
 477   * Only occurs at end of function.
 478
 479 1.6.6  SSG - Set Sign
 480
 481   dst.x = (src.x > 0.0) ? 1.0 : (src.x < 0.0) ? -1.0 : 0.0
 482   dst.y = (src.y > 0.0) ? 1.0 : (src.y < 0.0) ? -1.0 : 0.0
 483   dst.z = (src.z > 0.0) ? 1.0 : (src.z < 0.0) ? -1.0 : 0.0
 484   dst.w = (src.w > 0.0) ? 1.0 : (src.w < 0.0) ? -1.0 : 0.0
 485
 486
 487 1.8.1  CMP - Compare
 488
 489   dst.x = (src0.x < 0.0) ? src1.x : src2.x
 490   dst.y = (src0.y < 0.0) ? src1.y : src2.y
 491   dst.z = (src0.z < 0.0) ? src1.z : src2.z
 492   dst.w = (src0.w < 0.0) ? src1.w : src2.w
 493
 494
 495 1.8.2  KIL - Conditional Discard
 496
 497   if (src.x < 0.0 || src.y < 0.0 || src.z < 0.0 || src.w < 0.0)
 498     discard
 499   endif
 500
 501
 502 1.8.3  SCS - Sine Cosine
 503
 504   dst.x = cos(src.x)
 505   dst.y = sin(src.x)
 506   dst.z = 0.0
 507   dst.y = 1.0
 508
 509
 510 1.8.4  TXB - Texture Lookup With Bias
 511
 512   TBD
 513
 514
 515 1.9.1  NRM - 3-component Vector Normalise
 516
 517   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z)
 518   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z)
 519   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z)
 520   dst.w = 1.0
 521
 522
 523 1.9.2  DIV - Divide
 524
 525   dst.x = src0.x / src1.x
 526   dst.y = src0.y / src1.y
 527   dst.z = src0.z / src1.z
 528   dst.w = src0.w / src1.w
 529
 530
 531 1.9.3  DP2 - 2-component Dot Product
 532
 533   dst.x = src0.x * src1.x + src0.y * src1.y
 534   dst.y = src0.x * src1.x + src0.y * src1.y
 535   dst.z = src0.x * src1.x + src0.y * src1.y
 536   dst.w = src0.x * src1.x + src0.y * src1.y
 537
 538
 539 1.9.5  TXL - Texture Lookup With LOD
 540
 541   TBD
 542
 543
 544 1.9.6  BRK - Break
 545
 546   TBD
 547
 548
 549 1.9.7  IF - If
 550
 551   TBD
 552
 553
 554 1.9.8  BGNFOR - Begin a For-Loop
 555
 556   dst.x = floor(src.x)
 557   dst.y = floor(src.y)
 558   dst.z = floor(src.z)
 559
 560   if (dst.y <= 0)
 561     pc = [matching ENDFOR] + 1
 562   endif
 563
 564   Note: The destination must be a loop register.
 565         The source must be a constant register.
 566
 567   Considered for cleanup / removal.
 568
 569
 570 1.9.9  REP - Repeat
 571
 572   TBD
 573
 574
 575 1.9.10  ELSE - Else
 576
 577   TBD
 578
 579
 580 1.9.11  ENDIF - End If
 581
 582   TBD
 583
 584
 585 1.9.12  ENDFOR - End a For-Loop
 586
 587   dst.x = dst.x + dst.z
 588   dst.y = dst.y - 1.0
 589
 590   if (dst.y > 0)
 591     pc = [matching BGNFOR instruction] + 1
 592   endif
 593
 594   Note: The destination must be a loop register.
 595
 596   Considered for cleanup / removal.
 597
 598 1.9.13  ENDREP - End Repeat
 599
 600   TBD
 601
 602
 603 1.10.1  PUSHA - Push Address Register On Stack
 604
 605   push(src.x)
 606   push(src.y)
 607   push(src.z)
 608   push(src.w)
 609
 610   Considered for cleanup / removal.
 611
 612 1.10.2  POPA - Pop Address Register From Stack
 613
 614   dst.w = pop()
 615   dst.z = pop()
 616   dst.y = pop()
 617   dst.x = pop()
 618
 619   Considered for cleanup / removal.
 620
 621
 622 1.11  GL_NV_gpu_program4
 623 ------------------------
 624
 625 Support for these opcodes indicated by a special pipe capability bit (TBD).
 626
 627 1.11.1  CEIL - Ceiling
 628
 629   dst.x = ceil(src.x)
 630   dst.y = ceil(src.y)
 631   dst.z = ceil(src.z)
 632   dst.w = ceil(src.w)
 633
 634
 635 1.11.2  I2F - Integer To Float
 636
 637   dst.x = (float) src.x
 638   dst.y = (float) src.y
 639   dst.z = (float) src.z
 640   dst.w = (float) src.w
 641
 642
 643 1.11.3  NOT - Bitwise Not
 644
 645   dst.x = ~src.x
 646   dst.y = ~src.y
 647   dst.z = ~src.z
 648   dst.w = ~src.w
 649
 650
 651 1.11.4  TRUNC - Truncate
 652
 653   dst.x = trunc(src.x)
 654   dst.y = trunc(src.y)
 655   dst.z = trunc(src.z)
 656   dst.w = trunc(src.w)
 657
 658
 659 1.11.5  SHL - Shift Left
 660
 661   dst.x = src0.x << src1.x
 662   dst.y = src0.y << src1.x
 663   dst.z = src0.z << src1.x
 664   dst.w = src0.w << src1.x
 665
 666
 667 1.11.6  SHR - Shift Right
 668
 669   dst.x = src0.x >> src1.x
 670   dst.y = src0.y >> src1.x
 671   dst.z = src0.z >> src1.x
 672   dst.w = src0.w >> src1.x
 673
 674
 675 1.11.7  AND - Bitwise And
 676
 677   dst.x = src0.x & src1.x
 678   dst.y = src0.y & src1.y
 679   dst.z = src0.z & src1.z
 680   dst.w = src0.w & src1.w
 681
 682
 683 1.11.8  OR - Bitwise Or
 684
 685   dst.x = src0.x | src1.x
 686   dst.y = src0.y | src1.y
 687   dst.z = src0.z | src1.z
 688   dst.w = src0.w | src1.w
 689
 690
 691 1.11.9  MOD - Modulus
 692
 693   dst.x = src0.x % src1.x
 694   dst.y = src0.y % src1.y
 695   dst.z = src0.z % src1.z
 696   dst.w = src0.w % src1.w
 697
 698
 699 1.11.10  XOR - Bitwise Xor
 700
 701   dst.x = src0.x ^ src1.x
 702   dst.y = src0.y ^ src1.y
 703   dst.z = src0.z ^ src1.z
 704   dst.w = src0.w ^ src1.w
 705
 706
 707 1.11.11  SAD - Sum Of Absolute Differences
 708
 709   dst.x = abs(src0.x - src1.x) + src2.x
 710   dst.y = abs(src0.y - src1.y) + src2.y
 711   dst.z = abs(src0.z - src1.z) + src2.z
 712   dst.w = abs(src0.w - src1.w) + src2.w
 713
 714
 715 1.11.12  TXF - Texel Fetch
 716
 717   TBD
 718
 719
 720 1.11.13  TXQ - Texture Size Query
 721
 722   TBD
 723
 724
 725 1.11.14  CONT - Continue
 726
 727   TBD
 728
 729
 730 1.12  GL_NV_geometry_program4
 731 -----------------------------
 732
 733
 734 1.12.1  EMIT - Emit
 735
 736   TBD
 737
 738
 739 1.12.2  ENDPRIM - End Primitive
 740
 741   TBD
 742
 743
 744 1.13  GLSL
 745 ----------
 746
 747
 748 1.13.1  BGNLOOP - Begin a Loop
 749
 750   TBD
 751
 752
 753 1.13.2  BGNSUB - Begin Subroutine
 754
 755   TBD
 756
 757
 758 1.13.3  ENDLOOP - End a Loop
 759
 760   TBD
 761
 762
 763 1.13.4  ENDSUB - End Subroutine
 764
 765   TBD
 766
 767
 768
 769 1.13.10  NOP - No Operation
 770
 771   Do nothing.
 772
 773
 774
 775 1.16.7  NRM4 - 4-component Vector Normalise
 776
 777   dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 778   dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 779   dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 780   dst.w = src.w / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
 781
 782
 783 1.17  ps_2_x
 784 ------------
 785
 786
 787 1.17.2  CALLNZ - Subroutine Call If Not Zero
 788
 789   TBD
 790
 791
 792 1.17.3  IFC - If
 793
 794   TBD
 795
 796
 797 1.17.5  BREAKC - Break Conditional
 798
 799   TBD
 800
 801
 802 2  Explanation of symbols used
 803 ==============================
 804
 805
 806 2.1  Functions
 807 --------------
 808
 809
 810   abs(x)            Absolute value of x.
 811                     '|x|'
 812                     (x < 0.0) ? -x : x
 813
 814   ceil(x)           Ceiling of x.
 815
 816   clamp(x,y,z)      Clamp x between y and z.
 817                     (x < y) ? y : (x > z) ? z : x
 818
 819   cos(x)            Cosine of x.
 820
 821   floor(x)          Floor of x.
 822
 823   lg2(x)            Logarithm base 2 of x.
 824
 825   max(x,y)          Maximum of x and y.
 826                     (x > y) ? x : y
 827
 828   min(x,y)          Minimum of x and y.
 829                     (x < y) ? x : y
 830
 831   partialx(x)       Derivative of x relative to fragment's X.
 832
 833   partialy(x)       Derivative of x relative to fragment's Y.
 834
 835   pop()             Pop from stack.
 836
 837   pow(x,y)          Raise x to power of y.
 838
 839   push(x)           Push x on stack.
 840
 841   round(x)          Round x.
 842
 843   sin(x)            Sine of x.
 844
 845   sqrt(x)           Square root of x.
 846
 847   trunc(x)          Truncate x.
 848
 849
 850 2.2  Keywords
 851 -------------
 852
 853
 854   discard           Discard fragment.
 855
 856   dst               First destination register.
 857
 858   dst0              First destination register.
 859
 860   pc                Program counter.
 861
 862   src               First source register.
 863
 864   src0              First source register.
 865
 866   src1              Second source register.
 867
 868   src2              Third source register.
 869
 870   target            Label of target instruction.
 871
 872
 873 3  Other tokens
 874 ===============
 875
 876
 877 3.1  Declaration Semantic
 878 -------------------------
 879
 880
 881   Follows Declaration token if Semantic bit is set.
 882
 883   Since its purpose is to link a shader with other stages of the pipeline,
 884   it is valid to follow only those Declaration tokens that declare a register
 885   either in INPUT or OUTPUT file.
 886
 887   SemanticName field contains the semantic name of the register being declared.
 888   There is no default value.
 889
 890   SemanticIndex is an optional subscript that can be used to distinguish
 891   different register declarations with the same semantic name. The default value
 892   is 0.
 893
 894   The meanings of the individual semantic names are explained in the following
 895   sections.
 896
 897
 898 3.1.1  FACE
 899
 900   Valid only in a fragment shader INPUT declaration.
 901
 902   FACE.x is negative when the primitive is back facing. FACE.x is positive
 903   when the primitive is front facing.