4 TGSI, Tungsten Graphics Shader Infrastructure, 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
12 All TGSI instructions, known as *opcodes*, operate on arbitrary-precision
13 floating-point four-component vectors. An opcode may have up to one
14 destination register, known as *dst*, and between zero and three source
15 registers, called *src0* through *src2*, or simply *src* if there is only
18 Some instructions, like :opcode:`I2F`, permit re-interpretation of vector
19 components as integers. Other instructions permit using registers as
20 two-component vectors with double precision; see :ref:`doubleopcodes`.
22 When an instruction has a scalar result, the result is usually copied into
23 each of the components of *dst*. When this happens, the result is said to be
24 *replicated* to *dst*. :opcode:`RCP` is one such instruction.
29 TGSI supports modifiers on inputs (as well as saturate modifier on instructions).
31 For inputs which have a floating point type, both absolute value and negation
32 modifiers are supported (with absolute value being applied first).
33 TGSI_OPCODE_MOV is considered to have float input type for applying modifiers.
35 For inputs which have signed or unsigned type only the negate modifier is
42 ^^^^^^^^^^^^^^^^^^^^^^^^^
44 These opcodes are guaranteed to be available regardless of the driver being
47 .. opcode:: ARL - Address Register Load
51 dst.x = \lfloor src.x\rfloor
53 dst.y = \lfloor src.y\rfloor
55 dst.z = \lfloor src.z\rfloor
57 dst.w = \lfloor src.w\rfloor
60 .. opcode:: MOV - Move
73 .. opcode:: LIT - Light Coefficients
78 dst.y &= max(src.x, 0) \\
79 dst.z &= (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0 \\
83 .. opcode:: RCP - Reciprocal
85 This instruction replicates its result.
92 .. opcode:: RSQ - Reciprocal Square Root
94 This instruction replicates its result. The results are undefined for src <= 0.
98 dst = \frac{1}{\sqrt{src.x}}
101 .. opcode:: SQRT - Square Root
103 This instruction replicates its result. The results are undefined for src < 0.
110 .. opcode:: EXP - Approximate Exponential Base 2
114 dst.x &= 2^{\lfloor src.x\rfloor} \\
115 dst.y &= src.x - \lfloor src.x\rfloor \\
116 dst.z &= 2^{src.x} \\
120 .. opcode:: LOG - Approximate Logarithm Base 2
124 dst.x &= \lfloor\log_2{|src.x|}\rfloor \\
125 dst.y &= \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}} \\
126 dst.z &= \log_2{|src.x|} \\
130 .. opcode:: MUL - Multiply
134 dst.x = src0.x \times src1.x
136 dst.y = src0.y \times src1.y
138 dst.z = src0.z \times src1.z
140 dst.w = src0.w \times src1.w
143 .. opcode:: ADD - Add
147 dst.x = src0.x + src1.x
149 dst.y = src0.y + src1.y
151 dst.z = src0.z + src1.z
153 dst.w = src0.w + src1.w
156 .. opcode:: DP3 - 3-component Dot Product
158 This instruction replicates its result.
162 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z
165 .. opcode:: DP4 - 4-component Dot Product
167 This instruction replicates its result.
171 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src0.w \times src1.w
174 .. opcode:: DST - Distance Vector
179 dst.y &= src0.y \times src1.y\\
184 .. opcode:: MIN - Minimum
188 dst.x = min(src0.x, src1.x)
190 dst.y = min(src0.y, src1.y)
192 dst.z = min(src0.z, src1.z)
194 dst.w = min(src0.w, src1.w)
197 .. opcode:: MAX - Maximum
201 dst.x = max(src0.x, src1.x)
203 dst.y = max(src0.y, src1.y)
205 dst.z = max(src0.z, src1.z)
207 dst.w = max(src0.w, src1.w)
210 .. opcode:: SLT - Set On Less Than
214 dst.x = (src0.x < src1.x) ? 1.0F : 0.0F
216 dst.y = (src0.y < src1.y) ? 1.0F : 0.0F
218 dst.z = (src0.z < src1.z) ? 1.0F : 0.0F
220 dst.w = (src0.w < src1.w) ? 1.0F : 0.0F
223 .. opcode:: SGE - Set On Greater Equal Than
227 dst.x = (src0.x >= src1.x) ? 1.0F : 0.0F
229 dst.y = (src0.y >= src1.y) ? 1.0F : 0.0F
231 dst.z = (src0.z >= src1.z) ? 1.0F : 0.0F
233 dst.w = (src0.w >= src1.w) ? 1.0F : 0.0F
236 .. opcode:: MAD - Multiply And Add
240 dst.x = src0.x \times src1.x + src2.x
242 dst.y = src0.y \times src1.y + src2.y
244 dst.z = src0.z \times src1.z + src2.z
246 dst.w = src0.w \times src1.w + src2.w
249 .. opcode:: SUB - Subtract
253 dst.x = src0.x - src1.x
255 dst.y = src0.y - src1.y
257 dst.z = src0.z - src1.z
259 dst.w = src0.w - src1.w
262 .. opcode:: LRP - Linear Interpolate
266 dst.x = src0.x \times src1.x + (1 - src0.x) \times src2.x
268 dst.y = src0.y \times src1.y + (1 - src0.y) \times src2.y
270 dst.z = src0.z \times src1.z + (1 - src0.z) \times src2.z
272 dst.w = src0.w \times src1.w + (1 - src0.w) \times src2.w
275 .. opcode:: CND - Condition
279 dst.x = (src2.x > 0.5) ? src0.x : src1.x
281 dst.y = (src2.y > 0.5) ? src0.y : src1.y
283 dst.z = (src2.z > 0.5) ? src0.z : src1.z
285 dst.w = (src2.w > 0.5) ? src0.w : src1.w
288 .. opcode:: DP2A - 2-component Dot Product And Add
292 dst.x = src0.x \times src1.x + src0.y \times src1.y + src2.x
294 dst.y = src0.x \times src1.x + src0.y \times src1.y + src2.x
296 dst.z = src0.x \times src1.x + src0.y \times src1.y + src2.x
298 dst.w = src0.x \times src1.x + src0.y \times src1.y + src2.x
301 .. opcode:: FRC - Fraction
305 dst.x = src.x - \lfloor src.x\rfloor
307 dst.y = src.y - \lfloor src.y\rfloor
309 dst.z = src.z - \lfloor src.z\rfloor
311 dst.w = src.w - \lfloor src.w\rfloor
314 .. opcode:: CLAMP - Clamp
318 dst.x = clamp(src0.x, src1.x, src2.x)
320 dst.y = clamp(src0.y, src1.y, src2.y)
322 dst.z = clamp(src0.z, src1.z, src2.z)
324 dst.w = clamp(src0.w, src1.w, src2.w)
327 .. opcode:: FLR - Floor
329 This is identical to :opcode:`ARL`.
333 dst.x = \lfloor src.x\rfloor
335 dst.y = \lfloor src.y\rfloor
337 dst.z = \lfloor src.z\rfloor
339 dst.w = \lfloor src.w\rfloor
342 .. opcode:: ROUND - Round
355 .. opcode:: EX2 - Exponential Base 2
357 This instruction replicates its result.
364 .. opcode:: LG2 - Logarithm Base 2
366 This instruction replicates its result.
373 .. opcode:: POW - Power
375 This instruction replicates its result.
379 dst = src0.x^{src1.x}
381 .. opcode:: XPD - Cross Product
385 dst.x = src0.y \times src1.z - src1.y \times src0.z
387 dst.y = src0.z \times src1.x - src1.z \times src0.x
389 dst.z = src0.x \times src1.y - src1.x \times src0.y
394 .. opcode:: ABS - Absolute
407 .. opcode:: DPH - Homogeneous Dot Product
409 This instruction replicates its result.
413 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
416 .. opcode:: COS - Cosine
418 This instruction replicates its result.
425 .. opcode:: DDX, DDX_FINE - Derivative Relative To X
427 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
428 advertised. When it is, the fine version guarantees one derivative per row
429 while DDX is allowed to be the same for the entire 2x2 quad.
433 dst.x = partialx(src.x)
435 dst.y = partialx(src.y)
437 dst.z = partialx(src.z)
439 dst.w = partialx(src.w)
442 .. opcode:: DDY, DDY_FINE - Derivative Relative To Y
444 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
445 advertised. When it is, the fine version guarantees one derivative per column
446 while DDY is allowed to be the same for the entire 2x2 quad.
450 dst.x = partialy(src.x)
452 dst.y = partialy(src.y)
454 dst.z = partialy(src.z)
456 dst.w = partialy(src.w)
459 .. opcode:: PK2H - Pack Two 16-bit Floats
464 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
469 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
474 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
479 .. opcode:: SEQ - Set On Equal
483 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
485 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
487 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
489 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
492 .. opcode:: SGT - Set On Greater Than
496 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
498 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
500 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
502 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
505 .. opcode:: SIN - Sine
507 This instruction replicates its result.
514 .. opcode:: SLE - Set On Less Equal Than
518 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
520 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
522 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
524 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
527 .. opcode:: SNE - Set On Not Equal
531 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
533 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
535 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
537 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
540 .. opcode:: STR - Set On True
542 This instruction replicates its result.
549 .. opcode:: TEX - Texture Lookup
551 for array textures src0.y contains the slice for 1D,
552 and src0.z contain the slice for 2D.
554 for shadow textures with no arrays (and not cube map),
555 src0.z contains the reference value.
557 for shadow textures with arrays, src0.z contains
558 the reference value for 1D arrays, and src0.w contains
559 the reference value for 2D arrays and cube maps.
561 for cube map array shadow textures, the reference value
562 cannot be passed in src0.w, and TEX2 must be used instead.
568 shadow_ref = src0.z or src0.w (optional)
572 dst = texture\_sample(unit, coord, shadow_ref)
575 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
577 this is the same as TEX, but uses another reg to encode the
588 dst = texture\_sample(unit, coord, shadow_ref)
593 .. opcode:: TXD - Texture Lookup with Derivatives
605 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
608 .. opcode:: TXP - Projective Texture Lookup
612 coord.x = src0.x / src0.w
614 coord.y = src0.y / src0.w
616 coord.z = src0.z / src0.w
622 dst = texture\_sample(unit, coord)
625 .. opcode:: UP2H - Unpack Two 16-Bit Floats
631 Considered for removal.
633 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
639 Considered for removal.
641 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
647 Considered for removal.
649 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
655 Considered for removal.
658 .. opcode:: ARR - Address Register Load With Round
671 .. opcode:: SSG - Set Sign
675 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
677 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
679 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
681 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
684 .. opcode:: CMP - Compare
688 dst.x = (src0.x < 0) ? src1.x : src2.x
690 dst.y = (src0.y < 0) ? src1.y : src2.y
692 dst.z = (src0.z < 0) ? src1.z : src2.z
694 dst.w = (src0.w < 0) ? src1.w : src2.w
697 .. opcode:: KILL_IF - Conditional Discard
699 Conditional discard. Allowed in fragment shaders only.
703 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
708 .. opcode:: KILL - Discard
710 Unconditional discard. Allowed in fragment shaders only.
713 .. opcode:: SCS - Sine Cosine
726 .. opcode:: TXB - Texture Lookup With Bias
728 for cube map array textures and shadow cube maps, the bias value
729 cannot be passed in src0.w, and TXB2 must be used instead.
731 if the target is a shadow texture, the reference value is always
732 in src.z (this prevents shadow 3d and shadow 2d arrays from
733 using this instruction, but this is not needed).
749 dst = texture\_sample(unit, coord, bias)
752 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
754 this is the same as TXB, but uses another reg to encode the
755 lod bias value for cube map arrays and shadow cube maps.
756 Presumably shadow 2d arrays and shadow 3d targets could use
757 this encoding too, but this is not legal.
759 shadow cube map arrays are neither possible nor required.
769 dst = texture\_sample(unit, coord, bias)
772 .. opcode:: DIV - Divide
776 dst.x = \frac{src0.x}{src1.x}
778 dst.y = \frac{src0.y}{src1.y}
780 dst.z = \frac{src0.z}{src1.z}
782 dst.w = \frac{src0.w}{src1.w}
785 .. opcode:: DP2 - 2-component Dot Product
787 This instruction replicates its result.
791 dst = src0.x \times src1.x + src0.y \times src1.y
794 .. opcode:: TXL - Texture Lookup With explicit LOD
796 for cube map array textures, the explicit lod value
797 cannot be passed in src0.w, and TXL2 must be used instead.
799 if the target is a shadow texture, the reference value is always
800 in src.z (this prevents shadow 3d / 2d array / cube targets from
801 using this instruction, but this is not needed).
817 dst = texture\_sample(unit, coord, lod)
820 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
822 this is the same as TXL, but uses another reg to encode the
824 Presumably shadow 3d / 2d array / cube targets could use
825 this encoding too, but this is not legal.
827 shadow cube map arrays are neither possible nor required.
837 dst = texture\_sample(unit, coord, lod)
840 .. opcode:: PUSHA - Push Address Register On Stack
849 Considered for cleanup.
853 Considered for removal.
855 .. opcode:: POPA - Pop Address Register From Stack
864 Considered for cleanup.
868 Considered for removal.
871 .. opcode:: CALLNZ - Subroutine Call If Not Zero
877 Considered for cleanup.
881 Considered for removal.
885 ^^^^^^^^^^^^^^^^^^^^^^^^
887 These opcodes are primarily provided for special-use computational shaders.
888 Support for these opcodes indicated by a special pipe capability bit (TBD).
890 XXX doesn't look like most of the opcodes really belong here.
892 .. opcode:: CEIL - Ceiling
896 dst.x = \lceil src.x\rceil
898 dst.y = \lceil src.y\rceil
900 dst.z = \lceil src.z\rceil
902 dst.w = \lceil src.w\rceil
905 .. opcode:: TRUNC - Truncate
918 .. opcode:: MOD - Modulus
922 dst.x = src0.x \bmod src1.x
924 dst.y = src0.y \bmod src1.y
926 dst.z = src0.z \bmod src1.z
928 dst.w = src0.w \bmod src1.w
931 .. opcode:: UARL - Integer Address Register Load
933 Moves the contents of the source register, assumed to be an integer, into the
934 destination register, which is assumed to be an address (ADDR) register.
937 .. opcode:: SAD - Sum Of Absolute Differences
941 dst.x = |src0.x - src1.x| + src2.x
943 dst.y = |src0.y - src1.y| + src2.y
945 dst.z = |src0.z - src1.z| + src2.z
947 dst.w = |src0.w - src1.w| + src2.w
950 .. opcode:: TXF - Texel Fetch
952 As per NV_gpu_shader4, extract a single texel from a specified texture
953 image. The source sampler may not be a CUBE or SHADOW. src 0 is a
954 four-component signed integer vector used to identify the single texel
955 accessed. 3 components + level. Just like texture instructions, an optional
956 offset vector is provided, which is subject to various driver restrictions
957 (regarding range, source of offsets).
958 TXF(uint_vec coord, int_vec offset).
961 .. opcode:: TXQ - Texture Size Query
963 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
964 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
965 depth), 1D array (width, layers), 2D array (width, height, layers).
966 Also return the number of accessible levels (last_level - first_level + 1)
969 For components which don't return a resource dimension, their value
977 dst.x = texture\_width(unit, lod)
979 dst.y = texture\_height(unit, lod)
981 dst.z = texture\_depth(unit, lod)
983 dst.w = texture\_levels(unit)
985 .. opcode:: TG4 - Texture Gather
987 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
988 filtering operation and packs them into a single register. Only works with
989 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
990 addressing modes of the sampler and the top level of any mip pyramid are
991 used. Set W to zero. It behaves like the TEX instruction, but a filtered
992 sample is not generated. The four samples that contribute to filtering are
993 placed into xyzw in clockwise order, starting with the (u,v) texture
994 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
995 where the magnitude of the deltas are half a texel.
997 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
998 depth compares, single component selection, and a non-constant offset. It
999 doesn't allow support for the GL independent offset to get i0,j0. This would
1000 require another CAP is hw can do it natively. For now we lower that before
1009 dst = texture\_gather4 (unit, coord, component)
1011 (with SM5 - cube array shadow)
1019 dst = texture\_gather (uint, coord, compare)
1021 .. opcode:: LODQ - level of detail query
1023 Compute the LOD information that the texture pipe would use to access the
1024 texture. The Y component contains the computed LOD lambda_prime. The X
1025 component contains the LOD that will be accessed, based on min/max lod's
1032 dst.xy = lodq(uint, coord);
1035 ^^^^^^^^^^^^^^^^^^^^^^^^
1036 These opcodes are used for integer operations.
1037 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1040 .. opcode:: I2F - Signed Integer To Float
1042 Rounding is unspecified (round to nearest even suggested).
1046 dst.x = (float) src.x
1048 dst.y = (float) src.y
1050 dst.z = (float) src.z
1052 dst.w = (float) src.w
1055 .. opcode:: U2F - Unsigned Integer To Float
1057 Rounding is unspecified (round to nearest even suggested).
1061 dst.x = (float) src.x
1063 dst.y = (float) src.y
1065 dst.z = (float) src.z
1067 dst.w = (float) src.w
1070 .. opcode:: F2I - Float to Signed Integer
1072 Rounding is towards zero (truncate).
1073 Values outside signed range (including NaNs) produce undefined results.
1086 .. opcode:: F2U - Float to Unsigned Integer
1088 Rounding is towards zero (truncate).
1089 Values outside unsigned range (including NaNs) produce undefined results.
1093 dst.x = (unsigned) src.x
1095 dst.y = (unsigned) src.y
1097 dst.z = (unsigned) src.z
1099 dst.w = (unsigned) src.w
1102 .. opcode:: UADD - Integer Add
1104 This instruction works the same for signed and unsigned integers.
1105 The low 32bit of the result is returned.
1109 dst.x = src0.x + src1.x
1111 dst.y = src0.y + src1.y
1113 dst.z = src0.z + src1.z
1115 dst.w = src0.w + src1.w
1118 .. opcode:: UMAD - Integer Multiply And Add
1120 This instruction works the same for signed and unsigned integers.
1121 The multiplication returns the low 32bit (as does the result itself).
1125 dst.x = src0.x \times src1.x + src2.x
1127 dst.y = src0.y \times src1.y + src2.y
1129 dst.z = src0.z \times src1.z + src2.z
1131 dst.w = src0.w \times src1.w + src2.w
1134 .. opcode:: UMUL - Integer Multiply
1136 This instruction works the same for signed and unsigned integers.
1137 The low 32bit of the result is returned.
1141 dst.x = src0.x \times src1.x
1143 dst.y = src0.y \times src1.y
1145 dst.z = src0.z \times src1.z
1147 dst.w = src0.w \times src1.w
1150 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1152 The high 32bits of the multiplication of 2 signed integers are returned.
1156 dst.x = (src0.x \times src1.x) >> 32
1158 dst.y = (src0.y \times src1.y) >> 32
1160 dst.z = (src0.z \times src1.z) >> 32
1162 dst.w = (src0.w \times src1.w) >> 32
1165 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1167 The high 32bits of the multiplication of 2 unsigned integers are returned.
1171 dst.x = (src0.x \times src1.x) >> 32
1173 dst.y = (src0.y \times src1.y) >> 32
1175 dst.z = (src0.z \times src1.z) >> 32
1177 dst.w = (src0.w \times src1.w) >> 32
1180 .. opcode:: IDIV - Signed Integer Division
1182 TBD: behavior for division by zero.
1186 dst.x = src0.x \ src1.x
1188 dst.y = src0.y \ src1.y
1190 dst.z = src0.z \ src1.z
1192 dst.w = src0.w \ src1.w
1195 .. opcode:: UDIV - Unsigned Integer Division
1197 For division by zero, 0xffffffff is returned.
1201 dst.x = src0.x \ src1.x
1203 dst.y = src0.y \ src1.y
1205 dst.z = src0.z \ src1.z
1207 dst.w = src0.w \ src1.w
1210 .. opcode:: UMOD - Unsigned Integer Remainder
1212 If second arg is zero, 0xffffffff is returned.
1216 dst.x = src0.x \ src1.x
1218 dst.y = src0.y \ src1.y
1220 dst.z = src0.z \ src1.z
1222 dst.w = src0.w \ src1.w
1225 .. opcode:: NOT - Bitwise Not
1238 .. opcode:: AND - Bitwise And
1242 dst.x = src0.x \& src1.x
1244 dst.y = src0.y \& src1.y
1246 dst.z = src0.z \& src1.z
1248 dst.w = src0.w \& src1.w
1251 .. opcode:: OR - Bitwise Or
1255 dst.x = src0.x | src1.x
1257 dst.y = src0.y | src1.y
1259 dst.z = src0.z | src1.z
1261 dst.w = src0.w | src1.w
1264 .. opcode:: XOR - Bitwise Xor
1268 dst.x = src0.x \oplus src1.x
1270 dst.y = src0.y \oplus src1.y
1272 dst.z = src0.z \oplus src1.z
1274 dst.w = src0.w \oplus src1.w
1277 .. opcode:: IMAX - Maximum of Signed Integers
1281 dst.x = max(src0.x, src1.x)
1283 dst.y = max(src0.y, src1.y)
1285 dst.z = max(src0.z, src1.z)
1287 dst.w = max(src0.w, src1.w)
1290 .. opcode:: UMAX - Maximum of Unsigned Integers
1294 dst.x = max(src0.x, src1.x)
1296 dst.y = max(src0.y, src1.y)
1298 dst.z = max(src0.z, src1.z)
1300 dst.w = max(src0.w, src1.w)
1303 .. opcode:: IMIN - Minimum of Signed Integers
1307 dst.x = min(src0.x, src1.x)
1309 dst.y = min(src0.y, src1.y)
1311 dst.z = min(src0.z, src1.z)
1313 dst.w = min(src0.w, src1.w)
1316 .. opcode:: UMIN - Minimum of Unsigned Integers
1320 dst.x = min(src0.x, src1.x)
1322 dst.y = min(src0.y, src1.y)
1324 dst.z = min(src0.z, src1.z)
1326 dst.w = min(src0.w, src1.w)
1329 .. opcode:: SHL - Shift Left
1331 The shift count is masked with 0x1f before the shift is applied.
1335 dst.x = src0.x << (0x1f \& src1.x)
1337 dst.y = src0.y << (0x1f \& src1.y)
1339 dst.z = src0.z << (0x1f \& src1.z)
1341 dst.w = src0.w << (0x1f \& src1.w)
1344 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1346 The shift count is masked with 0x1f before the shift is applied.
1350 dst.x = src0.x >> (0x1f \& src1.x)
1352 dst.y = src0.y >> (0x1f \& src1.y)
1354 dst.z = src0.z >> (0x1f \& src1.z)
1356 dst.w = src0.w >> (0x1f \& src1.w)
1359 .. opcode:: USHR - Logical Shift Right
1361 The shift count is masked with 0x1f before the shift is applied.
1365 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1367 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1369 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1371 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1374 .. opcode:: UCMP - Integer Conditional Move
1378 dst.x = src0.x ? src1.x : src2.x
1380 dst.y = src0.y ? src1.y : src2.y
1382 dst.z = src0.z ? src1.z : src2.z
1384 dst.w = src0.w ? src1.w : src2.w
1388 .. opcode:: ISSG - Integer Set Sign
1392 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1394 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1396 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1398 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1402 .. opcode:: FSLT - Float Set On Less Than (ordered)
1404 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1408 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1410 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1412 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1414 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1417 .. opcode:: ISLT - Signed Integer Set On Less Than
1421 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1423 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1425 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1427 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1430 .. opcode:: USLT - Unsigned Integer Set On Less Than
1434 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1436 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1438 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1440 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1443 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1445 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1449 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1451 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1453 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1455 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1458 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1462 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1464 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1466 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1468 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1471 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1475 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1477 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1479 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1481 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1484 .. opcode:: FSEQ - Float Set On Equal (ordered)
1486 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1490 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1492 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1494 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1496 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1499 .. opcode:: USEQ - Integer Set On Equal
1503 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1505 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1507 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1509 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1512 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1514 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1518 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1520 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1522 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1524 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1527 .. opcode:: USNE - Integer Set On Not Equal
1531 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1533 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1535 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1537 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1540 .. opcode:: INEG - Integer Negate
1555 .. opcode:: IABS - Integer Absolute Value
1569 These opcodes are used for bit-level manipulation of integers.
1571 .. opcode:: IBFE - Signed Bitfield Extract
1573 See SM5 instruction of the same name. Extracts a set of bits from the input,
1574 and sign-extends them if the high bit of the extracted window is set.
1578 def ibfe(value, offset, bits):
1579 offset = offset & 0x1f
1581 if bits == 0: return 0
1582 # Note: >> sign-extends
1583 if width + offset < 32:
1584 return (value << (32 - offset - bits)) >> (32 - bits)
1586 return value >> offset
1588 .. opcode:: UBFE - Unsigned Bitfield Extract
1590 See SM5 instruction of the same name. Extracts a set of bits from the input,
1591 without any sign-extension.
1595 def ubfe(value, offset, bits):
1596 offset = offset & 0x1f
1598 if bits == 0: return 0
1599 # Note: >> does not sign-extend
1600 if width + offset < 32:
1601 return (value << (32 - offset - bits)) >> (32 - bits)
1603 return value >> offset
1605 .. opcode:: BFI - Bitfield Insert
1607 See SM5 instruction of the same name. Replaces a bit region of 'base' with
1608 the low bits of 'insert'.
1612 def bfi(base, insert, offset, bits):
1613 offset = offset & 0x1f
1615 mask = ((1 << bits) - 1) << offset
1616 return ((insert << offset) & mask) | (base & ~mask)
1618 .. opcode:: BREV - Bitfield Reverse
1620 See SM5 instruction BFREV. Reverses the bits of the argument.
1622 .. opcode:: POPC - Population Count
1624 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1626 .. opcode:: LSB - Index of lowest set bit
1628 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1629 bit of the argument. Returns -1 if none are set.
1631 .. opcode:: IMSB - Index of highest non-sign bit
1633 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1634 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1635 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1636 (i.e. for inputs 0 and -1).
1638 .. opcode:: UMSB - Index of highest set bit
1640 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1641 set bit of the argument. Returns -1 if none are set.
1644 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1646 These opcodes are only supported in geometry shaders; they have no meaning
1647 in any other type of shader.
1649 .. opcode:: EMIT - Emit
1651 Generate a new vertex for the current primitive into the specified vertex
1652 stream using the values in the output registers.
1655 .. opcode:: ENDPRIM - End Primitive
1657 Complete the current primitive in the specified vertex stream (consisting of
1658 the emitted vertices), and start a new one.
1664 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1665 opcodes is determined by a special capability bit, ``GLSL``.
1666 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1668 .. opcode:: CAL - Subroutine Call
1674 .. opcode:: RET - Subroutine Call Return
1679 .. opcode:: CONT - Continue
1681 Unconditionally moves the point of execution to the instruction after the
1682 last bgnloop. The instruction must appear within a bgnloop/endloop.
1686 Support for CONT is determined by a special capability bit,
1687 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1690 .. opcode:: BGNLOOP - Begin a Loop
1692 Start a loop. Must have a matching endloop.
1695 .. opcode:: BGNSUB - Begin Subroutine
1697 Starts definition of a subroutine. Must have a matching endsub.
1700 .. opcode:: ENDLOOP - End a Loop
1702 End a loop started with bgnloop.
1705 .. opcode:: ENDSUB - End Subroutine
1707 Ends definition of a subroutine.
1710 .. opcode:: NOP - No Operation
1715 .. opcode:: BRK - Break
1717 Unconditionally moves the point of execution to the instruction after the
1718 next endloop or endswitch. The instruction must appear within a loop/endloop
1719 or switch/endswitch.
1722 .. opcode:: BREAKC - Break Conditional
1724 Conditionally moves the point of execution to the instruction after the
1725 next endloop or endswitch. The instruction must appear within a loop/endloop
1726 or switch/endswitch.
1727 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1728 as an integer register.
1732 Considered for removal as it's quite inconsistent wrt other opcodes
1733 (could emulate with UIF/BRK/ENDIF).
1736 .. opcode:: IF - Float If
1738 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1742 where src0.x is interpreted as a floating point register.
1745 .. opcode:: UIF - Bitwise If
1747 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1751 where src0.x is interpreted as an integer register.
1754 .. opcode:: ELSE - Else
1756 Starts an else block, after an IF or UIF statement.
1759 .. opcode:: ENDIF - End If
1761 Ends an IF or UIF block.
1764 .. opcode:: SWITCH - Switch
1766 Starts a C-style switch expression. The switch consists of one or multiple
1767 CASE statements, and at most one DEFAULT statement. Execution of a statement
1768 ends when a BRK is hit, but just like in C falling through to other cases
1769 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1770 just as last statement, and fallthrough is allowed into/from it.
1771 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1777 (some instructions here)
1780 (some instructions here)
1783 (some instructions here)
1788 .. opcode:: CASE - Switch case
1790 This represents a switch case label. The src arg must be an integer immediate.
1793 .. opcode:: DEFAULT - Switch default
1795 This represents the default case in the switch, which is taken if no other
1799 .. opcode:: ENDSWITCH - End of switch
1801 Ends a switch expression.
1807 The interpolation instructions allow an input to be interpolated in a
1808 different way than its declaration. This corresponds to the GLSL 4.00
1809 interpolateAt* functions. The first argument of each of these must come from
1810 ``TGSI_FILE_INPUT``.
1812 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1814 Interpolates the varying specified by src0 at the centroid
1816 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1818 Interpolates the varying specified by src0 at the sample id specified by
1819 src1.x (interpreted as an integer)
1821 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1823 Interpolates the varying specified by src0 at the offset src1.xy from the
1824 pixel center (interpreted as floats)
1832 The double-precision opcodes reinterpret four-component vectors into
1833 two-component vectors with doubled precision in each component.
1835 Support for these opcodes is XXX undecided. :T
1837 .. opcode:: DADD - Add
1841 dst.xy = src0.xy + src1.xy
1843 dst.zw = src0.zw + src1.zw
1846 .. opcode:: DDIV - Divide
1850 dst.xy = src0.xy / src1.xy
1852 dst.zw = src0.zw / src1.zw
1854 .. opcode:: DSEQ - Set on Equal
1858 dst.xy = src0.xy == src1.xy ? 1.0F : 0.0F
1860 dst.zw = src0.zw == src1.zw ? 1.0F : 0.0F
1862 .. opcode:: DSLT - Set on Less than
1866 dst.xy = src0.xy < src1.xy ? 1.0F : 0.0F
1868 dst.zw = src0.zw < src1.zw ? 1.0F : 0.0F
1870 .. opcode:: DFRAC - Fraction
1874 dst.xy = src.xy - \lfloor src.xy\rfloor
1876 dst.zw = src.zw - \lfloor src.zw\rfloor
1879 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1881 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1882 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1883 :math:`dst1 \times 2^{dst0} = src` .
1887 dst0.xy = exp(src.xy)
1889 dst1.xy = frac(src.xy)
1891 dst0.zw = exp(src.zw)
1893 dst1.zw = frac(src.zw)
1895 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1897 This opcode is the inverse of :opcode:`DFRACEXP`.
1901 dst.xy = src0.xy \times 2^{src1.xy}
1903 dst.zw = src0.zw \times 2^{src1.zw}
1905 .. opcode:: DMIN - Minimum
1909 dst.xy = min(src0.xy, src1.xy)
1911 dst.zw = min(src0.zw, src1.zw)
1913 .. opcode:: DMAX - Maximum
1917 dst.xy = max(src0.xy, src1.xy)
1919 dst.zw = max(src0.zw, src1.zw)
1921 .. opcode:: DMUL - Multiply
1925 dst.xy = src0.xy \times src1.xy
1927 dst.zw = src0.zw \times src1.zw
1930 .. opcode:: DMAD - Multiply And Add
1934 dst.xy = src0.xy \times src1.xy + src2.xy
1936 dst.zw = src0.zw \times src1.zw + src2.zw
1939 .. opcode:: DRCP - Reciprocal
1943 dst.xy = \frac{1}{src.xy}
1945 dst.zw = \frac{1}{src.zw}
1947 .. opcode:: DSQRT - Square Root
1951 dst.xy = \sqrt{src.xy}
1953 dst.zw = \sqrt{src.zw}
1956 .. _samplingopcodes:
1958 Resource Sampling Opcodes
1959 ^^^^^^^^^^^^^^^^^^^^^^^^^
1961 Those opcodes follow very closely semantics of the respective Direct3D
1962 instructions. If in doubt double check Direct3D documentation.
1963 Note that the swizzle on SVIEW (src1) determines texel swizzling
1968 Using provided address, sample data from the specified texture using the
1969 filtering mode identified by the gven sampler. The source data may come from
1970 any resource type other than buffers.
1972 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
1974 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
1976 .. opcode:: SAMPLE_I
1978 Simplified alternative to the SAMPLE instruction. Using the provided
1979 integer address, SAMPLE_I fetches data from the specified sampler view
1980 without any filtering. The source data may come from any resource type
1983 Syntax: ``SAMPLE_I dst, address, sampler_view``
1985 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
1987 The 'address' is specified as unsigned integers. If the 'address' is out of
1988 range [0...(# texels - 1)] the result of the fetch is always 0 in all
1989 components. As such the instruction doesn't honor address wrap modes, in
1990 cases where that behavior is desirable 'SAMPLE' instruction should be used.
1991 address.w always provides an unsigned integer mipmap level. If the value is
1992 out of the range then the instruction always returns 0 in all components.
1993 address.yz are ignored for buffers and 1d textures. address.z is ignored
1994 for 1d texture arrays and 2d textures.
1996 For 1D texture arrays address.y provides the array index (also as unsigned
1997 integer). If the value is out of the range of available array indices
1998 [0... (array size - 1)] then the opcode always returns 0 in all components.
1999 For 2D texture arrays address.z provides the array index, otherwise it
2000 exhibits the same behavior as in the case for 1D texture arrays. The exact
2001 semantics of the source address are presented in the table below:
2003 +---------------------------+----+-----+-----+---------+
2004 | resource type | X | Y | Z | W |
2005 +===========================+====+=====+=====+=========+
2006 | ``PIPE_BUFFER`` | x | | | ignored |
2007 +---------------------------+----+-----+-----+---------+
2008 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2009 +---------------------------+----+-----+-----+---------+
2010 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2011 +---------------------------+----+-----+-----+---------+
2012 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2013 +---------------------------+----+-----+-----+---------+
2014 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2015 +---------------------------+----+-----+-----+---------+
2016 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2017 +---------------------------+----+-----+-----+---------+
2018 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2019 +---------------------------+----+-----+-----+---------+
2020 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2021 +---------------------------+----+-----+-----+---------+
2023 Where 'mpl' is a mipmap level and 'idx' is the array index.
2025 .. opcode:: SAMPLE_I_MS
2027 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2029 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2031 .. opcode:: SAMPLE_B
2033 Just like the SAMPLE instruction with the exception that an additional bias
2034 is applied to the level of detail computed as part of the instruction
2037 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2039 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2041 .. opcode:: SAMPLE_C
2043 Similar to the SAMPLE instruction but it performs a comparison filter. The
2044 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2045 additional float32 operand, reference value, which must be a register with
2046 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2047 current samplers compare_func (in pipe_sampler_state) to compare reference
2048 value against the red component value for the surce resource at each texel
2049 that the currently configured texture filter covers based on the provided
2052 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2054 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2056 .. opcode:: SAMPLE_C_LZ
2058 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2061 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2063 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2066 .. opcode:: SAMPLE_D
2068 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2069 the source address in the x direction and the y direction are provided by
2072 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2074 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2076 .. opcode:: SAMPLE_L
2078 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2079 directly as a scalar value, representing no anisotropy.
2081 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2083 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2087 Gathers the four texels to be used in a bi-linear filtering operation and
2088 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2089 and cubemaps arrays. For 2D textures, only the addressing modes of the
2090 sampler and the top level of any mip pyramid are used. Set W to zero. It
2091 behaves like the SAMPLE instruction, but a filtered sample is not
2092 generated. The four samples that contribute to filtering are placed into
2093 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2094 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2095 magnitude of the deltas are half a texel.
2098 .. opcode:: SVIEWINFO
2100 Query the dimensions of a given sampler view. dst receives width, height,
2101 depth or array size and number of mipmap levels as int4. The dst can have a
2102 writemask which will specify what info is the caller interested in.
2104 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2106 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2108 src_mip_level is an unsigned integer scalar. If it's out of range then
2109 returns 0 for width, height and depth/array size but the total number of
2110 mipmap is still returned correctly for the given sampler view. The returned
2111 width, height and depth values are for the mipmap level selected by the
2112 src_mip_level and are in the number of texels. For 1d texture array width
2113 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2114 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2115 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2116 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2117 resinfo allowing swizzling dst values is ignored (due to the interaction
2118 with rcpfloat modifier which requires some swizzle handling in the state
2121 .. opcode:: SAMPLE_POS
2123 Query the position of a given sample. dst receives float4 (x, y, 0, 0)
2124 indicated where the sample is located. If the resource is not a multi-sample
2125 resource and not a render target, the result is 0.
2127 .. opcode:: SAMPLE_INFO
2129 dst receives number of samples in x. If the resource is not a multi-sample
2130 resource and not a render target, the result is 0.
2133 .. _resourceopcodes:
2135 Resource Access Opcodes
2136 ^^^^^^^^^^^^^^^^^^^^^^^
2138 .. opcode:: LOAD - Fetch data from a shader resource
2140 Syntax: ``LOAD dst, resource, address``
2142 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
2144 Using the provided integer address, LOAD fetches data
2145 from the specified buffer or texture without any
2148 The 'address' is specified as a vector of unsigned
2149 integers. If the 'address' is out of range the result
2152 Only the first mipmap level of a resource can be read
2153 from using this instruction.
2155 For 1D or 2D texture arrays, the array index is
2156 provided as an unsigned integer in address.y or
2157 address.z, respectively. address.yz are ignored for
2158 buffers and 1D textures. address.z is ignored for 1D
2159 texture arrays and 2D textures. address.w is always
2162 .. opcode:: STORE - Write data to a shader resource
2164 Syntax: ``STORE resource, address, src``
2166 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2168 Using the provided integer address, STORE writes data
2169 to the specified buffer or texture.
2171 The 'address' is specified as a vector of unsigned
2172 integers. If the 'address' is out of range the result
2175 Only the first mipmap level of a resource can be
2176 written to using this instruction.
2178 For 1D or 2D texture arrays, the array index is
2179 provided as an unsigned integer in address.y or
2180 address.z, respectively. address.yz are ignored for
2181 buffers and 1D textures. address.z is ignored for 1D
2182 texture arrays and 2D textures. address.w is always
2186 .. _threadsyncopcodes:
2188 Inter-thread synchronization opcodes
2189 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2191 These opcodes are intended for communication between threads running
2192 within the same compute grid. For now they're only valid in compute
2195 .. opcode:: MFENCE - Memory fence
2197 Syntax: ``MFENCE resource``
2199 Example: ``MFENCE RES[0]``
2201 This opcode forces strong ordering between any memory access
2202 operations that affect the specified resource. This means that
2203 previous loads and stores (and only those) will be performed and
2204 visible to other threads before the program execution continues.
2207 .. opcode:: LFENCE - Load memory fence
2209 Syntax: ``LFENCE resource``
2211 Example: ``LFENCE RES[0]``
2213 Similar to MFENCE, but it only affects the ordering of memory loads.
2216 .. opcode:: SFENCE - Store memory fence
2218 Syntax: ``SFENCE resource``
2220 Example: ``SFENCE RES[0]``
2222 Similar to MFENCE, but it only affects the ordering of memory stores.
2225 .. opcode:: BARRIER - Thread group barrier
2229 This opcode suspends the execution of the current thread until all
2230 the remaining threads in the working group reach the same point of
2231 the program. Results are unspecified if any of the remaining
2232 threads terminates or never reaches an executed BARRIER instruction.
2240 These opcodes provide atomic variants of some common arithmetic and
2241 logical operations. In this context atomicity means that another
2242 concurrent memory access operation that affects the same memory
2243 location is guaranteed to be performed strictly before or after the
2244 entire execution of the atomic operation.
2246 For the moment they're only valid in compute programs.
2248 .. opcode:: ATOMUADD - Atomic integer addition
2250 Syntax: ``ATOMUADD dst, resource, offset, src``
2252 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2254 The following operation is performed atomically on each component:
2258 dst_i = resource[offset]_i
2260 resource[offset]_i = dst_i + src_i
2263 .. opcode:: ATOMXCHG - Atomic exchange
2265 Syntax: ``ATOMXCHG dst, resource, offset, src``
2267 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2269 The following operation is performed atomically on each component:
2273 dst_i = resource[offset]_i
2275 resource[offset]_i = src_i
2278 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2280 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2282 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2284 The following operation is performed atomically on each component:
2288 dst_i = resource[offset]_i
2290 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2293 .. opcode:: ATOMAND - Atomic bitwise And
2295 Syntax: ``ATOMAND dst, resource, offset, src``
2297 Example: ``ATOMAND TEMP[0], RES[0], TEMP[1], TEMP[2]``
2299 The following operation is performed atomically on each component:
2303 dst_i = resource[offset]_i
2305 resource[offset]_i = dst_i \& src_i
2308 .. opcode:: ATOMOR - Atomic bitwise Or
2310 Syntax: ``ATOMOR dst, resource, offset, src``
2312 Example: ``ATOMOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2314 The following operation is performed atomically on each component:
2318 dst_i = resource[offset]_i
2320 resource[offset]_i = dst_i | src_i
2323 .. opcode:: ATOMXOR - Atomic bitwise Xor
2325 Syntax: ``ATOMXOR dst, resource, offset, src``
2327 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2329 The following operation is performed atomically on each component:
2333 dst_i = resource[offset]_i
2335 resource[offset]_i = dst_i \oplus src_i
2338 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2340 Syntax: ``ATOMUMIN dst, resource, offset, src``
2342 Example: ``ATOMUMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2344 The following operation is performed atomically on each component:
2348 dst_i = resource[offset]_i
2350 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2353 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2355 Syntax: ``ATOMUMAX dst, resource, offset, src``
2357 Example: ``ATOMUMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2359 The following operation is performed atomically on each component:
2363 dst_i = resource[offset]_i
2365 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2368 .. opcode:: ATOMIMIN - Atomic signed minimum
2370 Syntax: ``ATOMIMIN dst, resource, offset, src``
2372 Example: ``ATOMIMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2374 The following operation is performed atomically on each component:
2378 dst_i = resource[offset]_i
2380 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2383 .. opcode:: ATOMIMAX - Atomic signed maximum
2385 Syntax: ``ATOMIMAX dst, resource, offset, src``
2387 Example: ``ATOMIMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2389 The following operation is performed atomically on each component:
2393 dst_i = resource[offset]_i
2395 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2399 Explanation of symbols used
2400 ------------------------------
2407 :math:`|x|` Absolute value of `x`.
2409 :math:`\lceil x \rceil` Ceiling of `x`.
2411 clamp(x,y,z) Clamp x between y and z.
2412 (x < y) ? y : (x > z) ? z : x
2414 :math:`\lfloor x\rfloor` Floor of `x`.
2416 :math:`\log_2{x}` Logarithm of `x`, base 2.
2418 max(x,y) Maximum of x and y.
2421 min(x,y) Minimum of x and y.
2424 partialx(x) Derivative of x relative to fragment's X.
2426 partialy(x) Derivative of x relative to fragment's Y.
2428 pop() Pop from stack.
2430 :math:`x^y` `x` to the power `y`.
2432 push(x) Push x on stack.
2436 trunc(x) Truncate x, i.e. drop the fraction bits.
2443 discard Discard fragment.
2447 target Label of target instruction.
2458 Declares a register that is will be referenced as an operand in Instruction
2461 File field contains register file that is being declared and is one
2464 UsageMask field specifies which of the register components can be accessed
2465 and is one of TGSI_WRITEMASK.
2467 The Local flag specifies that a given value isn't intended for
2468 subroutine parameter passing and, as a result, the implementation
2469 isn't required to give any guarantees of it being preserved across
2470 subroutine boundaries. As it's merely a compiler hint, the
2471 implementation is free to ignore it.
2473 If Dimension flag is set to 1, a Declaration Dimension token follows.
2475 If Semantic flag is set to 1, a Declaration Semantic token follows.
2477 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2479 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2481 If Array flag is set to 1, a Declaration Array token follows.
2484 ^^^^^^^^^^^^^^^^^^^^^^^^
2486 Declarations can optional have an ArrayID attribute which can be referred by
2487 indirect addressing operands. An ArrayID of zero is reserved and treaded as
2488 if no ArrayID is specified.
2490 If an indirect addressing operand refers to a specific declaration by using
2491 an ArrayID only the registers in this declaration are guaranteed to be
2492 accessed, accessing any register outside this declaration results in undefined
2493 behavior. Note that for compatibility the effective index is zero-based and
2494 not relative to the specified declaration
2496 If no ArrayID is specified with an indirect addressing operand the whole
2497 register file might be accessed by this operand. This is strongly discouraged
2498 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2500 Declaration Semantic
2501 ^^^^^^^^^^^^^^^^^^^^^^^^
2503 Vertex and fragment shader input and output registers may be labeled
2504 with semantic information consisting of a name and index.
2506 Follows Declaration token if Semantic bit is set.
2508 Since its purpose is to link a shader with other stages of the pipeline,
2509 it is valid to follow only those Declaration tokens that declare a register
2510 either in INPUT or OUTPUT file.
2512 SemanticName field contains the semantic name of the register being declared.
2513 There is no default value.
2515 SemanticIndex is an optional subscript that can be used to distinguish
2516 different register declarations with the same semantic name. The default value
2519 The meanings of the individual semantic names are explained in the following
2522 TGSI_SEMANTIC_POSITION
2523 """"""""""""""""""""""
2525 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2526 output register which contains the homogeneous vertex position in the clip
2527 space coordinate system. After clipping, the X, Y and Z components of the
2528 vertex will be divided by the W value to get normalized device coordinates.
2530 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2531 fragment shader input contains the fragment's window position. The X
2532 component starts at zero and always increases from left to right.
2533 The Y component starts at zero and always increases but Y=0 may either
2534 indicate the top of the window or the bottom depending on the fragment
2535 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2536 The Z coordinate ranges from 0 to 1 to represent depth from the front
2537 to the back of the Z buffer. The W component contains the reciprocol
2538 of the interpolated vertex position W component.
2540 Fragment shaders may also declare an output register with
2541 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2542 the fragment shader to change the fragment's Z position.
2549 For vertex shader outputs or fragment shader inputs/outputs, this
2550 label indicates that the resister contains an R,G,B,A color.
2552 Several shader inputs/outputs may contain colors so the semantic index
2553 is used to distinguish them. For example, color[0] may be the diffuse
2554 color while color[1] may be the specular color.
2556 This label is needed so that the flat/smooth shading can be applied
2557 to the right interpolants during rasterization.
2561 TGSI_SEMANTIC_BCOLOR
2562 """"""""""""""""""""
2564 Back-facing colors are only used for back-facing polygons, and are only valid
2565 in vertex shader outputs. After rasterization, all polygons are front-facing
2566 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2567 so all BCOLORs effectively become regular COLORs in the fragment shader.
2573 Vertex shader inputs and outputs and fragment shader inputs may be
2574 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2575 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2576 to compute a fog blend factor which is used to blend the normal fragment color
2577 with a constant fog color. But fog coord really is just an ordinary vec4
2578 register like regular semantics.
2584 Vertex shader input and output registers may be labeled with
2585 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2586 in the form (S, 0, 0, 1). The point size controls the width or diameter
2587 of points for rasterization. This label cannot be used in fragment
2590 When using this semantic, be sure to set the appropriate state in the
2591 :ref:`rasterizer` first.
2594 TGSI_SEMANTIC_TEXCOORD
2595 """"""""""""""""""""""
2597 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2599 Vertex shader outputs and fragment shader inputs may be labeled with
2600 this semantic to make them replaceable by sprite coordinates via the
2601 sprite_coord_enable state in the :ref:`rasterizer`.
2602 The semantic index permitted with this semantic is limited to <= 7.
2604 If the driver does not support TEXCOORD, sprite coordinate replacement
2605 applies to inputs with the GENERIC semantic instead.
2607 The intended use case for this semantic is gl_TexCoord.
2610 TGSI_SEMANTIC_PCOORD
2611 """"""""""""""""""""
2613 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2615 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2616 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2617 the current primitive is a point and point sprites are enabled. Otherwise,
2618 the contents of the register are undefined.
2620 The intended use case for this semantic is gl_PointCoord.
2623 TGSI_SEMANTIC_GENERIC
2624 """""""""""""""""""""
2626 All vertex/fragment shader inputs/outputs not labeled with any other
2627 semantic label can be considered to be generic attributes. Typical
2628 uses of generic inputs/outputs are texcoords and user-defined values.
2631 TGSI_SEMANTIC_NORMAL
2632 """"""""""""""""""""
2634 Indicates that a vertex shader input is a normal vector. This is
2635 typically only used for legacy graphics APIs.
2641 This label applies to fragment shader inputs only and indicates that
2642 the register contains front/back-face information of the form (F, 0,
2643 0, 1). The first component will be positive when the fragment belongs
2644 to a front-facing polygon, and negative when the fragment belongs to a
2645 back-facing polygon.
2648 TGSI_SEMANTIC_EDGEFLAG
2649 """"""""""""""""""""""
2651 For vertex shaders, this sematic label indicates that an input or
2652 output is a boolean edge flag. The register layout is [F, x, x, x]
2653 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2654 simply copies the edge flag input to the edgeflag output.
2656 Edge flags are used to control which lines or points are actually
2657 drawn when the polygon mode converts triangles/quads/polygons into
2661 TGSI_SEMANTIC_STENCIL
2662 """""""""""""""""""""
2664 For fragment shaders, this semantic label indicates that an output
2665 is a writable stencil reference value. Only the Y component is writable.
2666 This allows the fragment shader to change the fragments stencilref value.
2669 TGSI_SEMANTIC_VIEWPORT_INDEX
2670 """"""""""""""""""""""""""""
2672 For geometry shaders, this semantic label indicates that an output
2673 contains the index of the viewport (and scissor) to use.
2674 Only the X value is used.
2680 For geometry shaders, this semantic label indicates that an output
2681 contains the layer value to use for the color and depth/stencil surfaces.
2682 Only the X value is used. (Also known as rendertarget array index.)
2685 TGSI_SEMANTIC_CULLDIST
2686 """"""""""""""""""""""
2688 Used as distance to plane for performing application-defined culling
2689 of individual primitives against a plane. When components of vertex
2690 elements are given this label, these values are assumed to be a
2691 float32 signed distance to a plane. Primitives will be completely
2692 discarded if the plane distance for all of the vertices in the
2693 primitive are < 0. If a vertex has a cull distance of NaN, that
2694 vertex counts as "out" (as if its < 0);
2695 The limits on both clip and cull distances are bound
2696 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2697 the maximum number of components that can be used to hold the
2698 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2699 which specifies the maximum number of registers which can be
2700 annotated with those semantics.
2703 TGSI_SEMANTIC_CLIPDIST
2704 """"""""""""""""""""""
2706 When components of vertex elements are identified this way, these
2707 values are each assumed to be a float32 signed distance to a plane.
2708 Primitive setup only invokes rasterization on pixels for which
2709 the interpolated plane distances are >= 0. Multiple clip planes
2710 can be implemented simultaneously, by annotating multiple
2711 components of one or more vertex elements with the above specified
2712 semantic. The limits on both clip and cull distances are bound
2713 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2714 the maximum number of components that can be used to hold the
2715 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2716 which specifies the maximum number of registers which can be
2717 annotated with those semantics.
2719 TGSI_SEMANTIC_SAMPLEID
2720 """"""""""""""""""""""
2722 For fragment shaders, this semantic label indicates that a system value
2723 contains the current sample id (i.e. gl_SampleID). Only the X value is used.
2725 TGSI_SEMANTIC_SAMPLEPOS
2726 """""""""""""""""""""""
2728 For fragment shaders, this semantic label indicates that a system value
2729 contains the current sample's position (i.e. gl_SamplePosition). Only the X
2730 and Y values are used.
2732 TGSI_SEMANTIC_SAMPLEMASK
2733 """"""""""""""""""""""""
2735 For fragment shaders, this semantic label indicates that an output contains
2736 the sample mask used to disable further sample processing
2737 (i.e. gl_SampleMask). Only the X value is used, up to 32x MS.
2739 TGSI_SEMANTIC_INVOCATIONID
2740 """"""""""""""""""""""""""
2742 For geometry shaders, this semantic label indicates that a system value
2743 contains the current invocation id (i.e. gl_InvocationID). Only the X value is
2746 Declaration Interpolate
2747 ^^^^^^^^^^^^^^^^^^^^^^^
2749 This token is only valid for fragment shader INPUT declarations.
2751 The Interpolate field specifes the way input is being interpolated by
2752 the rasteriser and is one of TGSI_INTERPOLATE_*.
2754 The Location field specifies the location inside the pixel that the
2755 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
2756 when per-sample shading is enabled, the implementation may choose to
2757 interpolate at the sample irrespective of the Location field.
2759 The CylindricalWrap bitfield specifies which register components
2760 should be subject to cylindrical wrapping when interpolating by the
2761 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2762 should be interpolated according to cylindrical wrapping rules.
2765 Declaration Sampler View
2766 ^^^^^^^^^^^^^^^^^^^^^^^^
2768 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2770 DCL SVIEW[#], resource, type(s)
2772 Declares a shader input sampler view and assigns it to a SVIEW[#]
2775 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2777 type must be 1 or 4 entries (if specifying on a per-component
2778 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
2781 Declaration Resource
2782 ^^^^^^^^^^^^^^^^^^^^
2784 Follows Declaration token if file is TGSI_FILE_RESOURCE.
2786 DCL RES[#], resource [, WR] [, RAW]
2788 Declares a shader input resource and assigns it to a RES[#]
2791 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
2794 If the RAW keyword is not specified, the texture data will be
2795 subject to conversion, swizzling and scaling as required to yield
2796 the specified data type from the physical data format of the bound
2799 If the RAW keyword is specified, no channel conversion will be
2800 performed: the values read for each of the channels (X,Y,Z,W) will
2801 correspond to consecutive words in the same order and format
2802 they're found in memory. No element-to-address conversion will be
2803 performed either: the value of the provided X coordinate will be
2804 interpreted in byte units instead of texel units. The result of
2805 accessing a misaligned address is undefined.
2807 Usage of the STORE opcode is only allowed if the WR (writable) flag
2812 ^^^^^^^^^^^^^^^^^^^^^^^^
2814 Properties are general directives that apply to the whole TGSI program.
2819 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
2820 The default value is UPPER_LEFT.
2822 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
2823 increase downward and rightward.
2824 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
2825 increase upward and rightward.
2827 OpenGL defaults to LOWER_LEFT, and is configurable with the
2828 GL_ARB_fragment_coord_conventions extension.
2830 DirectX 9/10 use UPPER_LEFT.
2832 FS_COORD_PIXEL_CENTER
2833 """""""""""""""""""""
2835 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
2836 The default value is HALF_INTEGER.
2838 If HALF_INTEGER, the fractionary part of the position will be 0.5
2839 If INTEGER, the fractionary part of the position will be 0.0
2841 Note that this does not affect the set of fragments generated by
2842 rasterization, which is instead controlled by half_pixel_center in the
2845 OpenGL defaults to HALF_INTEGER, and is configurable with the
2846 GL_ARB_fragment_coord_conventions extension.
2848 DirectX 9 uses INTEGER.
2849 DirectX 10 uses HALF_INTEGER.
2851 FS_COLOR0_WRITES_ALL_CBUFS
2852 """"""""""""""""""""""""""
2853 Specifies that writes to the fragment shader color 0 are replicated to all
2854 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
2855 fragData is directed to a single color buffer, but fragColor is broadcast.
2858 """"""""""""""""""""""""""
2859 If this property is set on the program bound to the shader stage before the
2860 fragment shader, user clip planes should have no effect (be disabled) even if
2861 that shader does not write to any clip distance outputs and the rasterizer's
2862 clip_plane_enable is non-zero.
2863 This property is only supported by drivers that also support shader clip
2865 This is useful for APIs that don't have UCPs and where clip distances written
2866 by a shader cannot be disabled.
2871 Specifies the number of times a geometry shader should be executed for each
2872 input primitive. Each invocation will have a different
2873 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
2876 VS_WINDOW_SPACE_POSITION
2877 """"""""""""""""""""""""""
2878 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
2879 is assumed to contain window space coordinates.
2880 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
2881 directly taken from the 4-th component of the shader output.
2882 Naturally, clipping is not performed on window coordinates either.
2883 The effect of this property is undefined if a geometry or tessellation shader
2886 Texture Sampling and Texture Formats
2887 ------------------------------------
2889 This table shows how texture image components are returned as (x,y,z,w) tuples
2890 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
2891 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
2894 +--------------------+--------------+--------------------+--------------+
2895 | Texture Components | Gallium | OpenGL | Direct3D 9 |
2896 +====================+==============+====================+==============+
2897 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
2898 +--------------------+--------------+--------------------+--------------+
2899 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
2900 +--------------------+--------------+--------------------+--------------+
2901 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
2902 +--------------------+--------------+--------------------+--------------+
2903 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
2904 +--------------------+--------------+--------------------+--------------+
2905 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
2906 +--------------------+--------------+--------------------+--------------+
2907 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
2908 +--------------------+--------------+--------------------+--------------+
2909 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
2910 +--------------------+--------------+--------------------+--------------+
2911 | I | (i, i, i, i) | (i, i, i, i) | N/A |
2912 +--------------------+--------------+--------------------+--------------+
2913 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
2914 | | | [#envmap-bumpmap]_ | |
2915 +--------------------+--------------+--------------------+--------------+
2916 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
2917 | | | [#depth-tex-mode]_ | |
2918 +--------------------+--------------+--------------------+--------------+
2919 | S | (s, s, s, s) | unknown | unknown |
2920 +--------------------+--------------+--------------------+--------------+
2922 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
2923 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
2924 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.