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:: RCC - Reciprocal Clamped
409 This instruction replicates its result.
411 XXX cleanup on aisle three
415 dst = (1 / src.x) > 0 ? clamp(1 / src.x, 5.42101e-020, 1.84467e+019) : clamp(1 / src.x, -1.84467e+019, -5.42101e-020)
418 .. opcode:: DPH - Homogeneous Dot Product
420 This instruction replicates its result.
424 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
427 .. opcode:: COS - Cosine
429 This instruction replicates its result.
436 .. opcode:: DDX - Derivative Relative To X
440 dst.x = partialx(src.x)
442 dst.y = partialx(src.y)
444 dst.z = partialx(src.z)
446 dst.w = partialx(src.w)
449 .. opcode:: DDY - Derivative Relative To Y
453 dst.x = partialy(src.x)
455 dst.y = partialy(src.y)
457 dst.z = partialy(src.z)
459 dst.w = partialy(src.w)
462 .. opcode:: PK2H - Pack Two 16-bit Floats
467 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
472 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
477 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
482 .. opcode:: RFL - Reflection Vector
486 dst.x = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.x - src1.x
488 dst.y = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.y - src1.y
490 dst.z = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.z - src1.z
496 Considered for removal.
499 .. opcode:: SEQ - Set On Equal
503 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
505 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
507 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
509 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
512 .. opcode:: SFL - Set On False
514 This instruction replicates its result.
522 Considered for removal.
525 .. opcode:: SGT - Set On Greater Than
529 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
531 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
533 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
535 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
538 .. opcode:: SIN - Sine
540 This instruction replicates its result.
547 .. opcode:: SLE - Set On Less Equal Than
551 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
553 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
555 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
557 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
560 .. opcode:: SNE - Set On Not Equal
564 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
566 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
568 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
570 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
573 .. opcode:: STR - Set On True
575 This instruction replicates its result.
582 .. opcode:: TEX - Texture Lookup
584 for array textures src0.y contains the slice for 1D,
585 and src0.z contain the slice for 2D.
587 for shadow textures with no arrays (and not cube map),
588 src0.z contains the reference value.
590 for shadow textures with arrays, src0.z contains
591 the reference value for 1D arrays, and src0.w contains
592 the reference value for 2D arrays and cube maps.
594 for cube map array shadow textures, the reference value
595 cannot be passed in src0.w, and TEX2 must be used instead.
601 shadow_ref = src0.z or src0.w (optional)
605 dst = texture\_sample(unit, coord, shadow_ref)
608 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
610 this is the same as TEX, but uses another reg to encode the
621 dst = texture\_sample(unit, coord, shadow_ref)
626 .. opcode:: TXD - Texture Lookup with Derivatives
638 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
641 .. opcode:: TXP - Projective Texture Lookup
645 coord.x = src0.x / src0.w
647 coord.y = src0.y / src0.w
649 coord.z = src0.z / src0.w
655 dst = texture\_sample(unit, coord)
658 .. opcode:: UP2H - Unpack Two 16-Bit Floats
664 Considered for removal.
666 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
672 Considered for removal.
674 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
680 Considered for removal.
682 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
688 Considered for removal.
690 .. opcode:: X2D - 2D Coordinate Transformation
694 dst.x = src0.x + src1.x \times src2.x + src1.y \times src2.y
696 dst.y = src0.y + src1.x \times src2.z + src1.y \times src2.w
698 dst.z = src0.x + src1.x \times src2.x + src1.y \times src2.y
700 dst.w = src0.y + src1.x \times src2.z + src1.y \times src2.w
704 Considered for removal.
707 .. opcode:: ARA - Address Register Add
713 Considered for removal.
715 .. opcode:: ARR - Address Register Load With Round
728 .. opcode:: SSG - Set Sign
732 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
734 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
736 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
738 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
741 .. opcode:: CMP - Compare
745 dst.x = (src0.x < 0) ? src1.x : src2.x
747 dst.y = (src0.y < 0) ? src1.y : src2.y
749 dst.z = (src0.z < 0) ? src1.z : src2.z
751 dst.w = (src0.w < 0) ? src1.w : src2.w
754 .. opcode:: KILL_IF - Conditional Discard
756 Conditional discard. Allowed in fragment shaders only.
760 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
765 .. opcode:: KILL - Discard
767 Unconditional discard. Allowed in fragment shaders only.
770 .. opcode:: SCS - Sine Cosine
783 .. opcode:: TXB - Texture Lookup With Bias
785 for cube map array textures and shadow cube maps, the bias value
786 cannot be passed in src0.w, and TXB2 must be used instead.
788 if the target is a shadow texture, the reference value is always
789 in src.z (this prevents shadow 3d and shadow 2d arrays from
790 using this instruction, but this is not needed).
806 dst = texture\_sample(unit, coord, bias)
809 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
811 this is the same as TXB, but uses another reg to encode the
812 lod bias value for cube map arrays and shadow cube maps.
813 Presumably shadow 2d arrays and shadow 3d targets could use
814 this encoding too, but this is not legal.
816 shadow cube map arrays are neither possible nor required.
826 dst = texture\_sample(unit, coord, bias)
829 .. opcode:: NRM - 3-component Vector Normalise
833 dst.x = src.x / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
835 dst.y = src.y / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
837 dst.z = src.z / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
842 .. opcode:: DIV - Divide
846 dst.x = \frac{src0.x}{src1.x}
848 dst.y = \frac{src0.y}{src1.y}
850 dst.z = \frac{src0.z}{src1.z}
852 dst.w = \frac{src0.w}{src1.w}
855 .. opcode:: DP2 - 2-component Dot Product
857 This instruction replicates its result.
861 dst = src0.x \times src1.x + src0.y \times src1.y
864 .. opcode:: TXL - Texture Lookup With explicit LOD
866 for cube map array textures, the explicit lod value
867 cannot be passed in src0.w, and TXL2 must be used instead.
869 if the target is a shadow texture, the reference value is always
870 in src.z (this prevents shadow 3d / 2d array / cube targets from
871 using this instruction, but this is not needed).
887 dst = texture\_sample(unit, coord, lod)
890 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
892 this is the same as TXL, but uses another reg to encode the
894 Presumably shadow 3d / 2d array / cube targets could use
895 this encoding too, but this is not legal.
897 shadow cube map arrays are neither possible nor required.
907 dst = texture\_sample(unit, coord, lod)
910 .. opcode:: PUSHA - Push Address Register On Stack
919 Considered for cleanup.
923 Considered for removal.
925 .. opcode:: POPA - Pop Address Register From Stack
934 Considered for cleanup.
938 Considered for removal.
941 .. opcode:: BRA - Branch
947 Considered for removal.
950 .. opcode:: CALLNZ - Subroutine Call If Not Zero
956 Considered for cleanup.
960 Considered for removal.
964 ^^^^^^^^^^^^^^^^^^^^^^^^
966 These opcodes are primarily provided for special-use computational shaders.
967 Support for these opcodes indicated by a special pipe capability bit (TBD).
969 XXX doesn't look like most of the opcodes really belong here.
971 .. opcode:: CEIL - Ceiling
975 dst.x = \lceil src.x\rceil
977 dst.y = \lceil src.y\rceil
979 dst.z = \lceil src.z\rceil
981 dst.w = \lceil src.w\rceil
984 .. opcode:: TRUNC - Truncate
997 .. opcode:: MOD - Modulus
1001 dst.x = src0.x \bmod src1.x
1003 dst.y = src0.y \bmod src1.y
1005 dst.z = src0.z \bmod src1.z
1007 dst.w = src0.w \bmod src1.w
1010 .. opcode:: UARL - Integer Address Register Load
1012 Moves the contents of the source register, assumed to be an integer, into the
1013 destination register, which is assumed to be an address (ADDR) register.
1016 .. opcode:: SAD - Sum Of Absolute Differences
1020 dst.x = |src0.x - src1.x| + src2.x
1022 dst.y = |src0.y - src1.y| + src2.y
1024 dst.z = |src0.z - src1.z| + src2.z
1026 dst.w = |src0.w - src1.w| + src2.w
1029 .. opcode:: TXF - Texel Fetch
1031 As per NV_gpu_shader4, extract a single texel from a specified texture
1032 image. The source sampler may not be a CUBE or SHADOW. src 0 is a
1033 four-component signed integer vector used to identify the single texel
1034 accessed. 3 components + level. Just like texture instructions, an optional
1035 offset vector is provided, which is subject to various driver restrictions
1036 (regarding range, source of offsets).
1037 TXF(uint_vec coord, int_vec offset).
1040 .. opcode:: TXQ - Texture Size Query
1042 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
1043 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
1044 depth), 1D array (width, layers), 2D array (width, height, layers).
1045 Also return the number of accessible levels (last_level - first_level + 1)
1048 For components which don't return a resource dimension, their value
1056 dst.x = texture\_width(unit, lod)
1058 dst.y = texture\_height(unit, lod)
1060 dst.z = texture\_depth(unit, lod)
1062 dst.w = texture\_levels(unit)
1064 .. opcode:: TG4 - Texture Gather
1066 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
1067 filtering operation and packs them into a single register. Only works with
1068 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
1069 addressing modes of the sampler and the top level of any mip pyramid are
1070 used. Set W to zero. It behaves like the TEX instruction, but a filtered
1071 sample is not generated. The four samples that contribute to filtering are
1072 placed into xyzw in clockwise order, starting with the (u,v) texture
1073 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
1074 where the magnitude of the deltas are half a texel.
1076 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
1077 depth compares, single component selection, and a non-constant offset. It
1078 doesn't allow support for the GL independent offset to get i0,j0. This would
1079 require another CAP is hw can do it natively. For now we lower that before
1088 dst = texture\_gather4 (unit, coord, component)
1090 (with SM5 - cube array shadow)
1098 dst = texture\_gather (uint, coord, compare)
1100 .. opcode:: LODQ - level of detail query
1102 Compute the LOD information that the texture pipe would use to access the
1103 texture. The Y component contains the computed LOD lambda_prime. The X
1104 component contains the LOD that will be accessed, based on min/max lod's
1111 dst.xy = lodq(uint, coord);
1114 ^^^^^^^^^^^^^^^^^^^^^^^^
1115 These opcodes are used for integer operations.
1116 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1119 .. opcode:: I2F - Signed Integer To Float
1121 Rounding is unspecified (round to nearest even suggested).
1125 dst.x = (float) src.x
1127 dst.y = (float) src.y
1129 dst.z = (float) src.z
1131 dst.w = (float) src.w
1134 .. opcode:: U2F - Unsigned Integer To Float
1136 Rounding is unspecified (round to nearest even suggested).
1140 dst.x = (float) src.x
1142 dst.y = (float) src.y
1144 dst.z = (float) src.z
1146 dst.w = (float) src.w
1149 .. opcode:: F2I - Float to Signed Integer
1151 Rounding is towards zero (truncate).
1152 Values outside signed range (including NaNs) produce undefined results.
1165 .. opcode:: F2U - Float to Unsigned Integer
1167 Rounding is towards zero (truncate).
1168 Values outside unsigned range (including NaNs) produce undefined results.
1172 dst.x = (unsigned) src.x
1174 dst.y = (unsigned) src.y
1176 dst.z = (unsigned) src.z
1178 dst.w = (unsigned) src.w
1181 .. opcode:: UADD - Integer Add
1183 This instruction works the same for signed and unsigned integers.
1184 The low 32bit of the result is returned.
1188 dst.x = src0.x + src1.x
1190 dst.y = src0.y + src1.y
1192 dst.z = src0.z + src1.z
1194 dst.w = src0.w + src1.w
1197 .. opcode:: UMAD - Integer Multiply And Add
1199 This instruction works the same for signed and unsigned integers.
1200 The multiplication returns the low 32bit (as does the result itself).
1204 dst.x = src0.x \times src1.x + src2.x
1206 dst.y = src0.y \times src1.y + src2.y
1208 dst.z = src0.z \times src1.z + src2.z
1210 dst.w = src0.w \times src1.w + src2.w
1213 .. opcode:: UMUL - Integer Multiply
1215 This instruction works the same for signed and unsigned integers.
1216 The low 32bit of the result is returned.
1220 dst.x = src0.x \times src1.x
1222 dst.y = src0.y \times src1.y
1224 dst.z = src0.z \times src1.z
1226 dst.w = src0.w \times src1.w
1229 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1231 The high 32bits of the multiplication of 2 signed integers are returned.
1235 dst.x = (src0.x \times src1.x) >> 32
1237 dst.y = (src0.y \times src1.y) >> 32
1239 dst.z = (src0.z \times src1.z) >> 32
1241 dst.w = (src0.w \times src1.w) >> 32
1244 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1246 The high 32bits of the multiplication of 2 unsigned integers are returned.
1250 dst.x = (src0.x \times src1.x) >> 32
1252 dst.y = (src0.y \times src1.y) >> 32
1254 dst.z = (src0.z \times src1.z) >> 32
1256 dst.w = (src0.w \times src1.w) >> 32
1259 .. opcode:: IDIV - Signed Integer Division
1261 TBD: behavior for division by zero.
1265 dst.x = src0.x \ src1.x
1267 dst.y = src0.y \ src1.y
1269 dst.z = src0.z \ src1.z
1271 dst.w = src0.w \ src1.w
1274 .. opcode:: UDIV - Unsigned Integer Division
1276 For division by zero, 0xffffffff is returned.
1280 dst.x = src0.x \ src1.x
1282 dst.y = src0.y \ src1.y
1284 dst.z = src0.z \ src1.z
1286 dst.w = src0.w \ src1.w
1289 .. opcode:: UMOD - Unsigned Integer Remainder
1291 If second arg is zero, 0xffffffff is returned.
1295 dst.x = src0.x \ src1.x
1297 dst.y = src0.y \ src1.y
1299 dst.z = src0.z \ src1.z
1301 dst.w = src0.w \ src1.w
1304 .. opcode:: NOT - Bitwise Not
1317 .. opcode:: AND - Bitwise And
1321 dst.x = src0.x \& src1.x
1323 dst.y = src0.y \& src1.y
1325 dst.z = src0.z \& src1.z
1327 dst.w = src0.w \& src1.w
1330 .. opcode:: OR - Bitwise Or
1334 dst.x = src0.x | src1.x
1336 dst.y = src0.y | src1.y
1338 dst.z = src0.z | src1.z
1340 dst.w = src0.w | src1.w
1343 .. opcode:: XOR - Bitwise Xor
1347 dst.x = src0.x \oplus src1.x
1349 dst.y = src0.y \oplus src1.y
1351 dst.z = src0.z \oplus src1.z
1353 dst.w = src0.w \oplus src1.w
1356 .. opcode:: IMAX - Maximum of Signed Integers
1360 dst.x = max(src0.x, src1.x)
1362 dst.y = max(src0.y, src1.y)
1364 dst.z = max(src0.z, src1.z)
1366 dst.w = max(src0.w, src1.w)
1369 .. opcode:: UMAX - Maximum of Unsigned Integers
1373 dst.x = max(src0.x, src1.x)
1375 dst.y = max(src0.y, src1.y)
1377 dst.z = max(src0.z, src1.z)
1379 dst.w = max(src0.w, src1.w)
1382 .. opcode:: IMIN - Minimum of Signed Integers
1386 dst.x = min(src0.x, src1.x)
1388 dst.y = min(src0.y, src1.y)
1390 dst.z = min(src0.z, src1.z)
1392 dst.w = min(src0.w, src1.w)
1395 .. opcode:: UMIN - Minimum of Unsigned Integers
1399 dst.x = min(src0.x, src1.x)
1401 dst.y = min(src0.y, src1.y)
1403 dst.z = min(src0.z, src1.z)
1405 dst.w = min(src0.w, src1.w)
1408 .. opcode:: SHL - Shift Left
1410 The shift count is masked with 0x1f before the shift is applied.
1414 dst.x = src0.x << (0x1f \& src1.x)
1416 dst.y = src0.y << (0x1f \& src1.y)
1418 dst.z = src0.z << (0x1f \& src1.z)
1420 dst.w = src0.w << (0x1f \& src1.w)
1423 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1425 The shift count is masked with 0x1f before the shift is applied.
1429 dst.x = src0.x >> (0x1f \& src1.x)
1431 dst.y = src0.y >> (0x1f \& src1.y)
1433 dst.z = src0.z >> (0x1f \& src1.z)
1435 dst.w = src0.w >> (0x1f \& src1.w)
1438 .. opcode:: USHR - Logical Shift Right
1440 The shift count is masked with 0x1f before the shift is applied.
1444 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1446 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1448 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1450 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1453 .. opcode:: UCMP - Integer Conditional Move
1457 dst.x = src0.x ? src1.x : src2.x
1459 dst.y = src0.y ? src1.y : src2.y
1461 dst.z = src0.z ? src1.z : src2.z
1463 dst.w = src0.w ? src1.w : src2.w
1467 .. opcode:: ISSG - Integer Set Sign
1471 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1473 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1475 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1477 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1481 .. opcode:: FSLT - Float Set On Less Than (ordered)
1483 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1487 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1489 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1491 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1493 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1496 .. opcode:: ISLT - Signed Integer Set On Less Than
1500 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1502 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1504 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1506 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1509 .. opcode:: USLT - Unsigned Integer Set On Less Than
1513 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1515 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1517 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1519 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1522 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1524 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1528 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1530 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1532 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1534 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1537 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1541 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1543 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1545 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1547 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1550 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1554 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1556 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1558 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1560 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1563 .. opcode:: FSEQ - Float Set On Equal (ordered)
1565 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1569 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1571 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1573 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1575 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1578 .. opcode:: USEQ - Integer Set On Equal
1582 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1584 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1586 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1588 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1591 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1593 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1597 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1599 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1601 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1603 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1606 .. opcode:: USNE - Integer Set On Not Equal
1610 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1612 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1614 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1616 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1619 .. opcode:: INEG - Integer Negate
1634 .. opcode:: IABS - Integer Absolute Value
1648 These opcodes are used for bit-level manipulation of integers.
1650 .. opcode:: IBFE - Signed Bitfield Extract
1652 See SM5 instruction of the same name. Extracts a set of bits from the input,
1653 and sign-extends them if the high bit of the extracted window is set.
1657 def ibfe(value, offset, bits):
1658 offset = offset & 0x1f
1660 if bits == 0: return 0
1661 # Note: >> sign-extends
1662 if width + offset < 32:
1663 return (value << (32 - offset - bits)) >> (32 - bits)
1665 return value >> offset
1667 .. opcode:: UBFE - Unsigned Bitfield Extract
1669 See SM5 instruction of the same name. Extracts a set of bits from the input,
1670 without any sign-extension.
1674 def ubfe(value, offset, bits):
1675 offset = offset & 0x1f
1677 if bits == 0: return 0
1678 # Note: >> does not sign-extend
1679 if width + offset < 32:
1680 return (value << (32 - offset - bits)) >> (32 - bits)
1682 return value >> offset
1684 .. opcode:: BFI - Bitfield Insert
1686 See SM5 instruction of the same name. Replaces a bit region of 'base' with
1687 the low bits of 'insert'.
1691 def bfi(base, insert, offset, bits):
1692 offset = offset & 0x1f
1694 mask = ((1 << bits) - 1) << offset
1695 return ((insert << offset) & mask) | (base & ~mask)
1697 .. opcode:: BREV - Bitfield Reverse
1699 See SM5 instruction BFREV. Reverses the bits of the argument.
1701 .. opcode:: POPC - Population Count
1703 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1705 .. opcode:: LSB - Index of lowest set bit
1707 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1708 bit of the argument. Returns -1 if none are set.
1710 .. opcode:: IMSB - Index of highest non-sign bit
1712 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1713 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1714 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1715 (i.e. for inputs 0 and -1).
1717 .. opcode:: UMSB - Index of highest set bit
1719 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1720 set bit of the argument. Returns -1 if none are set.
1723 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1725 These opcodes are only supported in geometry shaders; they have no meaning
1726 in any other type of shader.
1728 .. opcode:: EMIT - Emit
1730 Generate a new vertex for the current primitive into the specified vertex
1731 stream using the values in the output registers.
1734 .. opcode:: ENDPRIM - End Primitive
1736 Complete the current primitive in the specified vertex stream (consisting of
1737 the emitted vertices), and start a new one.
1743 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1744 opcodes is determined by a special capability bit, ``GLSL``.
1745 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1747 .. opcode:: CAL - Subroutine Call
1753 .. opcode:: RET - Subroutine Call Return
1758 .. opcode:: CONT - Continue
1760 Unconditionally moves the point of execution to the instruction after the
1761 last bgnloop. The instruction must appear within a bgnloop/endloop.
1765 Support for CONT is determined by a special capability bit,
1766 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1769 .. opcode:: BGNLOOP - Begin a Loop
1771 Start a loop. Must have a matching endloop.
1774 .. opcode:: BGNSUB - Begin Subroutine
1776 Starts definition of a subroutine. Must have a matching endsub.
1779 .. opcode:: ENDLOOP - End a Loop
1781 End a loop started with bgnloop.
1784 .. opcode:: ENDSUB - End Subroutine
1786 Ends definition of a subroutine.
1789 .. opcode:: NOP - No Operation
1794 .. opcode:: BRK - Break
1796 Unconditionally moves the point of execution to the instruction after the
1797 next endloop or endswitch. The instruction must appear within a loop/endloop
1798 or switch/endswitch.
1801 .. opcode:: BREAKC - Break Conditional
1803 Conditionally moves the point of execution to the instruction after the
1804 next endloop or endswitch. The instruction must appear within a loop/endloop
1805 or switch/endswitch.
1806 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1807 as an integer register.
1811 Considered for removal as it's quite inconsistent wrt other opcodes
1812 (could emulate with UIF/BRK/ENDIF).
1815 .. opcode:: IF - Float If
1817 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1821 where src0.x is interpreted as a floating point register.
1824 .. opcode:: UIF - Bitwise If
1826 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1830 where src0.x is interpreted as an integer register.
1833 .. opcode:: ELSE - Else
1835 Starts an else block, after an IF or UIF statement.
1838 .. opcode:: ENDIF - End If
1840 Ends an IF or UIF block.
1843 .. opcode:: SWITCH - Switch
1845 Starts a C-style switch expression. The switch consists of one or multiple
1846 CASE statements, and at most one DEFAULT statement. Execution of a statement
1847 ends when a BRK is hit, but just like in C falling through to other cases
1848 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1849 just as last statement, and fallthrough is allowed into/from it.
1850 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1856 (some instructions here)
1859 (some instructions here)
1862 (some instructions here)
1867 .. opcode:: CASE - Switch case
1869 This represents a switch case label. The src arg must be an integer immediate.
1872 .. opcode:: DEFAULT - Switch default
1874 This represents the default case in the switch, which is taken if no other
1878 .. opcode:: ENDSWITCH - End of switch
1880 Ends a switch expression.
1883 .. opcode:: NRM4 - 4-component Vector Normalise
1885 This instruction replicates its result.
1889 dst = \frac{src.x}{src.x \times src.x + src.y \times src.y + src.z \times src.z + src.w \times src.w}
1895 The interpolation instructions allow an input to be interpolated in a
1896 different way than its declaration. This corresponds to the GLSL 4.00
1897 interpolateAt* functions. The first argument of each of these must come from
1898 ``TGSI_FILE_INPUT``.
1900 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1902 Interpolates the varying specified by src0 at the centroid
1904 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1906 Interpolates the varying specified by src0 at the sample id specified by
1907 src1.x (interpreted as an integer)
1909 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1911 Interpolates the varying specified by src0 at the offset src1.xy from the
1912 pixel center (interpreted as floats)
1920 The double-precision opcodes reinterpret four-component vectors into
1921 two-component vectors with doubled precision in each component.
1923 Support for these opcodes is XXX undecided. :T
1925 .. opcode:: DADD - Add
1929 dst.xy = src0.xy + src1.xy
1931 dst.zw = src0.zw + src1.zw
1934 .. opcode:: DDIV - Divide
1938 dst.xy = src0.xy / src1.xy
1940 dst.zw = src0.zw / src1.zw
1942 .. opcode:: DSEQ - Set on Equal
1946 dst.xy = src0.xy == src1.xy ? 1.0F : 0.0F
1948 dst.zw = src0.zw == src1.zw ? 1.0F : 0.0F
1950 .. opcode:: DSLT - Set on Less than
1954 dst.xy = src0.xy < src1.xy ? 1.0F : 0.0F
1956 dst.zw = src0.zw < src1.zw ? 1.0F : 0.0F
1958 .. opcode:: DFRAC - Fraction
1962 dst.xy = src.xy - \lfloor src.xy\rfloor
1964 dst.zw = src.zw - \lfloor src.zw\rfloor
1967 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1969 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1970 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1971 :math:`dst1 \times 2^{dst0} = src` .
1975 dst0.xy = exp(src.xy)
1977 dst1.xy = frac(src.xy)
1979 dst0.zw = exp(src.zw)
1981 dst1.zw = frac(src.zw)
1983 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1985 This opcode is the inverse of :opcode:`DFRACEXP`.
1989 dst.xy = src0.xy \times 2^{src1.xy}
1991 dst.zw = src0.zw \times 2^{src1.zw}
1993 .. opcode:: DMIN - Minimum
1997 dst.xy = min(src0.xy, src1.xy)
1999 dst.zw = min(src0.zw, src1.zw)
2001 .. opcode:: DMAX - Maximum
2005 dst.xy = max(src0.xy, src1.xy)
2007 dst.zw = max(src0.zw, src1.zw)
2009 .. opcode:: DMUL - Multiply
2013 dst.xy = src0.xy \times src1.xy
2015 dst.zw = src0.zw \times src1.zw
2018 .. opcode:: DMAD - Multiply And Add
2022 dst.xy = src0.xy \times src1.xy + src2.xy
2024 dst.zw = src0.zw \times src1.zw + src2.zw
2027 .. opcode:: DRCP - Reciprocal
2031 dst.xy = \frac{1}{src.xy}
2033 dst.zw = \frac{1}{src.zw}
2035 .. opcode:: DSQRT - Square Root
2039 dst.xy = \sqrt{src.xy}
2041 dst.zw = \sqrt{src.zw}
2044 .. _samplingopcodes:
2046 Resource Sampling Opcodes
2047 ^^^^^^^^^^^^^^^^^^^^^^^^^
2049 Those opcodes follow very closely semantics of the respective Direct3D
2050 instructions. If in doubt double check Direct3D documentation.
2051 Note that the swizzle on SVIEW (src1) determines texel swizzling
2056 Using provided address, sample data from the specified texture using the
2057 filtering mode identified by the gven sampler. The source data may come from
2058 any resource type other than buffers.
2060 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
2062 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
2064 .. opcode:: SAMPLE_I
2066 Simplified alternative to the SAMPLE instruction. Using the provided
2067 integer address, SAMPLE_I fetches data from the specified sampler view
2068 without any filtering. The source data may come from any resource type
2071 Syntax: ``SAMPLE_I dst, address, sampler_view``
2073 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
2075 The 'address' is specified as unsigned integers. If the 'address' is out of
2076 range [0...(# texels - 1)] the result of the fetch is always 0 in all
2077 components. As such the instruction doesn't honor address wrap modes, in
2078 cases where that behavior is desirable 'SAMPLE' instruction should be used.
2079 address.w always provides an unsigned integer mipmap level. If the value is
2080 out of the range then the instruction always returns 0 in all components.
2081 address.yz are ignored for buffers and 1d textures. address.z is ignored
2082 for 1d texture arrays and 2d textures.
2084 For 1D texture arrays address.y provides the array index (also as unsigned
2085 integer). If the value is out of the range of available array indices
2086 [0... (array size - 1)] then the opcode always returns 0 in all components.
2087 For 2D texture arrays address.z provides the array index, otherwise it
2088 exhibits the same behavior as in the case for 1D texture arrays. The exact
2089 semantics of the source address are presented in the table below:
2091 +---------------------------+----+-----+-----+---------+
2092 | resource type | X | Y | Z | W |
2093 +===========================+====+=====+=====+=========+
2094 | ``PIPE_BUFFER`` | x | | | ignored |
2095 +---------------------------+----+-----+-----+---------+
2096 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2097 +---------------------------+----+-----+-----+---------+
2098 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2099 +---------------------------+----+-----+-----+---------+
2100 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2101 +---------------------------+----+-----+-----+---------+
2102 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2103 +---------------------------+----+-----+-----+---------+
2104 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2105 +---------------------------+----+-----+-----+---------+
2106 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2107 +---------------------------+----+-----+-----+---------+
2108 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2109 +---------------------------+----+-----+-----+---------+
2111 Where 'mpl' is a mipmap level and 'idx' is the array index.
2113 .. opcode:: SAMPLE_I_MS
2115 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2117 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2119 .. opcode:: SAMPLE_B
2121 Just like the SAMPLE instruction with the exception that an additional bias
2122 is applied to the level of detail computed as part of the instruction
2125 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2127 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2129 .. opcode:: SAMPLE_C
2131 Similar to the SAMPLE instruction but it performs a comparison filter. The
2132 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2133 additional float32 operand, reference value, which must be a register with
2134 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2135 current samplers compare_func (in pipe_sampler_state) to compare reference
2136 value against the red component value for the surce resource at each texel
2137 that the currently configured texture filter covers based on the provided
2140 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2142 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2144 .. opcode:: SAMPLE_C_LZ
2146 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2149 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2151 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2154 .. opcode:: SAMPLE_D
2156 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2157 the source address in the x direction and the y direction are provided by
2160 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2162 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2164 .. opcode:: SAMPLE_L
2166 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2167 directly as a scalar value, representing no anisotropy.
2169 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2171 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2175 Gathers the four texels to be used in a bi-linear filtering operation and
2176 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2177 and cubemaps arrays. For 2D textures, only the addressing modes of the
2178 sampler and the top level of any mip pyramid are used. Set W to zero. It
2179 behaves like the SAMPLE instruction, but a filtered sample is not
2180 generated. The four samples that contribute to filtering are placed into
2181 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2182 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2183 magnitude of the deltas are half a texel.
2186 .. opcode:: SVIEWINFO
2188 Query the dimensions of a given sampler view. dst receives width, height,
2189 depth or array size and number of mipmap levels as int4. The dst can have a
2190 writemask which will specify what info is the caller interested in.
2192 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2194 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2196 src_mip_level is an unsigned integer scalar. If it's out of range then
2197 returns 0 for width, height and depth/array size but the total number of
2198 mipmap is still returned correctly for the given sampler view. The returned
2199 width, height and depth values are for the mipmap level selected by the
2200 src_mip_level and are in the number of texels. For 1d texture array width
2201 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2202 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2203 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2204 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2205 resinfo allowing swizzling dst values is ignored (due to the interaction
2206 with rcpfloat modifier which requires some swizzle handling in the state
2209 .. opcode:: SAMPLE_POS
2211 Query the position of a given sample. dst receives float4 (x, y, 0, 0)
2212 indicated where the sample is located. If the resource is not a multi-sample
2213 resource and not a render target, the result is 0.
2215 .. opcode:: SAMPLE_INFO
2217 dst receives number of samples in x. If the resource is not a multi-sample
2218 resource and not a render target, the result is 0.
2221 .. _resourceopcodes:
2223 Resource Access Opcodes
2224 ^^^^^^^^^^^^^^^^^^^^^^^
2226 .. opcode:: LOAD - Fetch data from a shader resource
2228 Syntax: ``LOAD dst, resource, address``
2230 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
2232 Using the provided integer address, LOAD fetches data
2233 from the specified buffer or texture without any
2236 The 'address' is specified as a vector of unsigned
2237 integers. If the 'address' is out of range the result
2240 Only the first mipmap level of a resource can be read
2241 from using this instruction.
2243 For 1D or 2D texture arrays, the array index is
2244 provided as an unsigned integer in address.y or
2245 address.z, respectively. address.yz are ignored for
2246 buffers and 1D textures. address.z is ignored for 1D
2247 texture arrays and 2D textures. address.w is always
2250 .. opcode:: STORE - Write data to a shader resource
2252 Syntax: ``STORE resource, address, src``
2254 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2256 Using the provided integer address, STORE writes data
2257 to the specified buffer or texture.
2259 The 'address' is specified as a vector of unsigned
2260 integers. If the 'address' is out of range the result
2263 Only the first mipmap level of a resource can be
2264 written to using this instruction.
2266 For 1D or 2D texture arrays, the array index is
2267 provided as an unsigned integer in address.y or
2268 address.z, respectively. address.yz are ignored for
2269 buffers and 1D textures. address.z is ignored for 1D
2270 texture arrays and 2D textures. address.w is always
2274 .. _threadsyncopcodes:
2276 Inter-thread synchronization opcodes
2277 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2279 These opcodes are intended for communication between threads running
2280 within the same compute grid. For now they're only valid in compute
2283 .. opcode:: MFENCE - Memory fence
2285 Syntax: ``MFENCE resource``
2287 Example: ``MFENCE RES[0]``
2289 This opcode forces strong ordering between any memory access
2290 operations that affect the specified resource. This means that
2291 previous loads and stores (and only those) will be performed and
2292 visible to other threads before the program execution continues.
2295 .. opcode:: LFENCE - Load memory fence
2297 Syntax: ``LFENCE resource``
2299 Example: ``LFENCE RES[0]``
2301 Similar to MFENCE, but it only affects the ordering of memory loads.
2304 .. opcode:: SFENCE - Store memory fence
2306 Syntax: ``SFENCE resource``
2308 Example: ``SFENCE RES[0]``
2310 Similar to MFENCE, but it only affects the ordering of memory stores.
2313 .. opcode:: BARRIER - Thread group barrier
2317 This opcode suspends the execution of the current thread until all
2318 the remaining threads in the working group reach the same point of
2319 the program. Results are unspecified if any of the remaining
2320 threads terminates or never reaches an executed BARRIER instruction.
2328 These opcodes provide atomic variants of some common arithmetic and
2329 logical operations. In this context atomicity means that another
2330 concurrent memory access operation that affects the same memory
2331 location is guaranteed to be performed strictly before or after the
2332 entire execution of the atomic operation.
2334 For the moment they're only valid in compute programs.
2336 .. opcode:: ATOMUADD - Atomic integer addition
2338 Syntax: ``ATOMUADD dst, resource, offset, src``
2340 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2342 The following operation is performed atomically on each component:
2346 dst_i = resource[offset]_i
2348 resource[offset]_i = dst_i + src_i
2351 .. opcode:: ATOMXCHG - Atomic exchange
2353 Syntax: ``ATOMXCHG dst, resource, offset, src``
2355 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2357 The following operation is performed atomically on each component:
2361 dst_i = resource[offset]_i
2363 resource[offset]_i = src_i
2366 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2368 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2370 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2372 The following operation is performed atomically on each component:
2376 dst_i = resource[offset]_i
2378 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2381 .. opcode:: ATOMAND - Atomic bitwise And
2383 Syntax: ``ATOMAND dst, resource, offset, src``
2385 Example: ``ATOMAND TEMP[0], RES[0], TEMP[1], TEMP[2]``
2387 The following operation is performed atomically on each component:
2391 dst_i = resource[offset]_i
2393 resource[offset]_i = dst_i \& src_i
2396 .. opcode:: ATOMOR - Atomic bitwise Or
2398 Syntax: ``ATOMOR dst, resource, offset, src``
2400 Example: ``ATOMOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2402 The following operation is performed atomically on each component:
2406 dst_i = resource[offset]_i
2408 resource[offset]_i = dst_i | src_i
2411 .. opcode:: ATOMXOR - Atomic bitwise Xor
2413 Syntax: ``ATOMXOR dst, resource, offset, src``
2415 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2417 The following operation is performed atomically on each component:
2421 dst_i = resource[offset]_i
2423 resource[offset]_i = dst_i \oplus src_i
2426 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2428 Syntax: ``ATOMUMIN dst, resource, offset, src``
2430 Example: ``ATOMUMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2432 The following operation is performed atomically on each component:
2436 dst_i = resource[offset]_i
2438 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2441 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2443 Syntax: ``ATOMUMAX dst, resource, offset, src``
2445 Example: ``ATOMUMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2447 The following operation is performed atomically on each component:
2451 dst_i = resource[offset]_i
2453 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2456 .. opcode:: ATOMIMIN - Atomic signed minimum
2458 Syntax: ``ATOMIMIN dst, resource, offset, src``
2460 Example: ``ATOMIMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2462 The following operation is performed atomically on each component:
2466 dst_i = resource[offset]_i
2468 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2471 .. opcode:: ATOMIMAX - Atomic signed maximum
2473 Syntax: ``ATOMIMAX dst, resource, offset, src``
2475 Example: ``ATOMIMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2477 The following operation is performed atomically on each component:
2481 dst_i = resource[offset]_i
2483 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2487 Explanation of symbols used
2488 ------------------------------
2495 :math:`|x|` Absolute value of `x`.
2497 :math:`\lceil x \rceil` Ceiling of `x`.
2499 clamp(x,y,z) Clamp x between y and z.
2500 (x < y) ? y : (x > z) ? z : x
2502 :math:`\lfloor x\rfloor` Floor of `x`.
2504 :math:`\log_2{x}` Logarithm of `x`, base 2.
2506 max(x,y) Maximum of x and y.
2509 min(x,y) Minimum of x and y.
2512 partialx(x) Derivative of x relative to fragment's X.
2514 partialy(x) Derivative of x relative to fragment's Y.
2516 pop() Pop from stack.
2518 :math:`x^y` `x` to the power `y`.
2520 push(x) Push x on stack.
2524 trunc(x) Truncate x, i.e. drop the fraction bits.
2531 discard Discard fragment.
2535 target Label of target instruction.
2546 Declares a register that is will be referenced as an operand in Instruction
2549 File field contains register file that is being declared and is one
2552 UsageMask field specifies which of the register components can be accessed
2553 and is one of TGSI_WRITEMASK.
2555 The Local flag specifies that a given value isn't intended for
2556 subroutine parameter passing and, as a result, the implementation
2557 isn't required to give any guarantees of it being preserved across
2558 subroutine boundaries. As it's merely a compiler hint, the
2559 implementation is free to ignore it.
2561 If Dimension flag is set to 1, a Declaration Dimension token follows.
2563 If Semantic flag is set to 1, a Declaration Semantic token follows.
2565 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2567 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2569 If Array flag is set to 1, a Declaration Array token follows.
2572 ^^^^^^^^^^^^^^^^^^^^^^^^
2574 Declarations can optional have an ArrayID attribute which can be referred by
2575 indirect addressing operands. An ArrayID of zero is reserved and treaded as
2576 if no ArrayID is specified.
2578 If an indirect addressing operand refers to a specific declaration by using
2579 an ArrayID only the registers in this declaration are guaranteed to be
2580 accessed, accessing any register outside this declaration results in undefined
2581 behavior. Note that for compatibility the effective index is zero-based and
2582 not relative to the specified declaration
2584 If no ArrayID is specified with an indirect addressing operand the whole
2585 register file might be accessed by this operand. This is strongly discouraged
2586 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2588 Declaration Semantic
2589 ^^^^^^^^^^^^^^^^^^^^^^^^
2591 Vertex and fragment shader input and output registers may be labeled
2592 with semantic information consisting of a name and index.
2594 Follows Declaration token if Semantic bit is set.
2596 Since its purpose is to link a shader with other stages of the pipeline,
2597 it is valid to follow only those Declaration tokens that declare a register
2598 either in INPUT or OUTPUT file.
2600 SemanticName field contains the semantic name of the register being declared.
2601 There is no default value.
2603 SemanticIndex is an optional subscript that can be used to distinguish
2604 different register declarations with the same semantic name. The default value
2607 The meanings of the individual semantic names are explained in the following
2610 TGSI_SEMANTIC_POSITION
2611 """"""""""""""""""""""
2613 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2614 output register which contains the homogeneous vertex position in the clip
2615 space coordinate system. After clipping, the X, Y and Z components of the
2616 vertex will be divided by the W value to get normalized device coordinates.
2618 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2619 fragment shader input contains the fragment's window position. The X
2620 component starts at zero and always increases from left to right.
2621 The Y component starts at zero and always increases but Y=0 may either
2622 indicate the top of the window or the bottom depending on the fragment
2623 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2624 The Z coordinate ranges from 0 to 1 to represent depth from the front
2625 to the back of the Z buffer. The W component contains the reciprocol
2626 of the interpolated vertex position W component.
2628 Fragment shaders may also declare an output register with
2629 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2630 the fragment shader to change the fragment's Z position.
2637 For vertex shader outputs or fragment shader inputs/outputs, this
2638 label indicates that the resister contains an R,G,B,A color.
2640 Several shader inputs/outputs may contain colors so the semantic index
2641 is used to distinguish them. For example, color[0] may be the diffuse
2642 color while color[1] may be the specular color.
2644 This label is needed so that the flat/smooth shading can be applied
2645 to the right interpolants during rasterization.
2649 TGSI_SEMANTIC_BCOLOR
2650 """"""""""""""""""""
2652 Back-facing colors are only used for back-facing polygons, and are only valid
2653 in vertex shader outputs. After rasterization, all polygons are front-facing
2654 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2655 so all BCOLORs effectively become regular COLORs in the fragment shader.
2661 Vertex shader inputs and outputs and fragment shader inputs may be
2662 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2663 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2664 to compute a fog blend factor which is used to blend the normal fragment color
2665 with a constant fog color. But fog coord really is just an ordinary vec4
2666 register like regular semantics.
2672 Vertex shader input and output registers may be labeled with
2673 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2674 in the form (S, 0, 0, 1). The point size controls the width or diameter
2675 of points for rasterization. This label cannot be used in fragment
2678 When using this semantic, be sure to set the appropriate state in the
2679 :ref:`rasterizer` first.
2682 TGSI_SEMANTIC_TEXCOORD
2683 """"""""""""""""""""""
2685 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2687 Vertex shader outputs and fragment shader inputs may be labeled with
2688 this semantic to make them replaceable by sprite coordinates via the
2689 sprite_coord_enable state in the :ref:`rasterizer`.
2690 The semantic index permitted with this semantic is limited to <= 7.
2692 If the driver does not support TEXCOORD, sprite coordinate replacement
2693 applies to inputs with the GENERIC semantic instead.
2695 The intended use case for this semantic is gl_TexCoord.
2698 TGSI_SEMANTIC_PCOORD
2699 """"""""""""""""""""
2701 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2703 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2704 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2705 the current primitive is a point and point sprites are enabled. Otherwise,
2706 the contents of the register are undefined.
2708 The intended use case for this semantic is gl_PointCoord.
2711 TGSI_SEMANTIC_GENERIC
2712 """""""""""""""""""""
2714 All vertex/fragment shader inputs/outputs not labeled with any other
2715 semantic label can be considered to be generic attributes. Typical
2716 uses of generic inputs/outputs are texcoords and user-defined values.
2719 TGSI_SEMANTIC_NORMAL
2720 """"""""""""""""""""
2722 Indicates that a vertex shader input is a normal vector. This is
2723 typically only used for legacy graphics APIs.
2729 This label applies to fragment shader inputs only and indicates that
2730 the register contains front/back-face information of the form (F, 0,
2731 0, 1). The first component will be positive when the fragment belongs
2732 to a front-facing polygon, and negative when the fragment belongs to a
2733 back-facing polygon.
2736 TGSI_SEMANTIC_EDGEFLAG
2737 """"""""""""""""""""""
2739 For vertex shaders, this sematic label indicates that an input or
2740 output is a boolean edge flag. The register layout is [F, x, x, x]
2741 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2742 simply copies the edge flag input to the edgeflag output.
2744 Edge flags are used to control which lines or points are actually
2745 drawn when the polygon mode converts triangles/quads/polygons into
2749 TGSI_SEMANTIC_STENCIL
2750 """""""""""""""""""""
2752 For fragment shaders, this semantic label indicates that an output
2753 is a writable stencil reference value. Only the Y component is writable.
2754 This allows the fragment shader to change the fragments stencilref value.
2757 TGSI_SEMANTIC_VIEWPORT_INDEX
2758 """"""""""""""""""""""""""""
2760 For geometry shaders, this semantic label indicates that an output
2761 contains the index of the viewport (and scissor) to use.
2762 Only the X value is used.
2768 For geometry shaders, this semantic label indicates that an output
2769 contains the layer value to use for the color and depth/stencil surfaces.
2770 Only the X value is used. (Also known as rendertarget array index.)
2773 TGSI_SEMANTIC_CULLDIST
2774 """"""""""""""""""""""
2776 Used as distance to plane for performing application-defined culling
2777 of individual primitives against a plane. When components of vertex
2778 elements are given this label, these values are assumed to be a
2779 float32 signed distance to a plane. Primitives will be completely
2780 discarded if the plane distance for all of the vertices in the
2781 primitive are < 0. If a vertex has a cull distance of NaN, that
2782 vertex counts as "out" (as if its < 0);
2783 The limits on both clip and cull distances are bound
2784 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2785 the maximum number of components that can be used to hold the
2786 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2787 which specifies the maximum number of registers which can be
2788 annotated with those semantics.
2791 TGSI_SEMANTIC_CLIPDIST
2792 """"""""""""""""""""""
2794 When components of vertex elements are identified this way, these
2795 values are each assumed to be a float32 signed distance to a plane.
2796 Primitive setup only invokes rasterization on pixels for which
2797 the interpolated plane distances are >= 0. Multiple clip planes
2798 can be implemented simultaneously, by annotating multiple
2799 components of one or more vertex elements with the above specified
2800 semantic. The limits on both clip and cull distances are bound
2801 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2802 the maximum number of components that can be used to hold the
2803 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2804 which specifies the maximum number of registers which can be
2805 annotated with those semantics.
2807 TGSI_SEMANTIC_SAMPLEID
2808 """"""""""""""""""""""
2810 For fragment shaders, this semantic label indicates that a system value
2811 contains the current sample id (i.e. gl_SampleID). Only the X value is used.
2813 TGSI_SEMANTIC_SAMPLEPOS
2814 """""""""""""""""""""""
2816 For fragment shaders, this semantic label indicates that a system value
2817 contains the current sample's position (i.e. gl_SamplePosition). Only the X
2818 and Y values are used.
2820 TGSI_SEMANTIC_SAMPLEMASK
2821 """"""""""""""""""""""""
2823 For fragment shaders, this semantic label indicates that an output contains
2824 the sample mask used to disable further sample processing
2825 (i.e. gl_SampleMask). Only the X value is used, up to 32x MS.
2827 TGSI_SEMANTIC_INVOCATIONID
2828 """"""""""""""""""""""""""
2830 For geometry shaders, this semantic label indicates that a system value
2831 contains the current invocation id (i.e. gl_InvocationID). Only the X value is
2834 Declaration Interpolate
2835 ^^^^^^^^^^^^^^^^^^^^^^^
2837 This token is only valid for fragment shader INPUT declarations.
2839 The Interpolate field specifes the way input is being interpolated by
2840 the rasteriser and is one of TGSI_INTERPOLATE_*.
2842 The Location field specifies the location inside the pixel that the
2843 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
2844 when per-sample shading is enabled, the implementation may choose to
2845 interpolate at the sample irrespective of the Location field.
2847 The CylindricalWrap bitfield specifies which register components
2848 should be subject to cylindrical wrapping when interpolating by the
2849 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2850 should be interpolated according to cylindrical wrapping rules.
2853 Declaration Sampler View
2854 ^^^^^^^^^^^^^^^^^^^^^^^^
2856 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2858 DCL SVIEW[#], resource, type(s)
2860 Declares a shader input sampler view and assigns it to a SVIEW[#]
2863 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2865 type must be 1 or 4 entries (if specifying on a per-component
2866 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
2869 Declaration Resource
2870 ^^^^^^^^^^^^^^^^^^^^
2872 Follows Declaration token if file is TGSI_FILE_RESOURCE.
2874 DCL RES[#], resource [, WR] [, RAW]
2876 Declares a shader input resource and assigns it to a RES[#]
2879 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
2882 If the RAW keyword is not specified, the texture data will be
2883 subject to conversion, swizzling and scaling as required to yield
2884 the specified data type from the physical data format of the bound
2887 If the RAW keyword is specified, no channel conversion will be
2888 performed: the values read for each of the channels (X,Y,Z,W) will
2889 correspond to consecutive words in the same order and format
2890 they're found in memory. No element-to-address conversion will be
2891 performed either: the value of the provided X coordinate will be
2892 interpreted in byte units instead of texel units. The result of
2893 accessing a misaligned address is undefined.
2895 Usage of the STORE opcode is only allowed if the WR (writable) flag
2900 ^^^^^^^^^^^^^^^^^^^^^^^^
2902 Properties are general directives that apply to the whole TGSI program.
2907 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
2908 The default value is UPPER_LEFT.
2910 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
2911 increase downward and rightward.
2912 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
2913 increase upward and rightward.
2915 OpenGL defaults to LOWER_LEFT, and is configurable with the
2916 GL_ARB_fragment_coord_conventions extension.
2918 DirectX 9/10 use UPPER_LEFT.
2920 FS_COORD_PIXEL_CENTER
2921 """""""""""""""""""""
2923 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
2924 The default value is HALF_INTEGER.
2926 If HALF_INTEGER, the fractionary part of the position will be 0.5
2927 If INTEGER, the fractionary part of the position will be 0.0
2929 Note that this does not affect the set of fragments generated by
2930 rasterization, which is instead controlled by half_pixel_center in the
2933 OpenGL defaults to HALF_INTEGER, and is configurable with the
2934 GL_ARB_fragment_coord_conventions extension.
2936 DirectX 9 uses INTEGER.
2937 DirectX 10 uses HALF_INTEGER.
2939 FS_COLOR0_WRITES_ALL_CBUFS
2940 """"""""""""""""""""""""""
2941 Specifies that writes to the fragment shader color 0 are replicated to all
2942 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
2943 fragData is directed to a single color buffer, but fragColor is broadcast.
2946 """"""""""""""""""""""""""
2947 If this property is set on the program bound to the shader stage before the
2948 fragment shader, user clip planes should have no effect (be disabled) even if
2949 that shader does not write to any clip distance outputs and the rasterizer's
2950 clip_plane_enable is non-zero.
2951 This property is only supported by drivers that also support shader clip
2953 This is useful for APIs that don't have UCPs and where clip distances written
2954 by a shader cannot be disabled.
2959 Specifies the number of times a geometry shader should be executed for each
2960 input primitive. Each invocation will have a different
2961 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
2964 VS_WINDOW_SPACE_POSITION
2965 """"""""""""""""""""""""""
2966 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
2967 is assumed to contain window space coordinates.
2968 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
2969 directly taken from the 4-th component of the shader output.
2970 Naturally, clipping is not performed on window coordinates either.
2971 The effect of this property is undefined if a geometry or tessellation shader
2974 Texture Sampling and Texture Formats
2975 ------------------------------------
2977 This table shows how texture image components are returned as (x,y,z,w) tuples
2978 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
2979 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
2982 +--------------------+--------------+--------------------+--------------+
2983 | Texture Components | Gallium | OpenGL | Direct3D 9 |
2984 +====================+==============+====================+==============+
2985 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
2986 +--------------------+--------------+--------------------+--------------+
2987 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
2988 +--------------------+--------------+--------------------+--------------+
2989 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
2990 +--------------------+--------------+--------------------+--------------+
2991 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
2992 +--------------------+--------------+--------------------+--------------+
2993 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
2994 +--------------------+--------------+--------------------+--------------+
2995 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
2996 +--------------------+--------------+--------------------+--------------+
2997 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
2998 +--------------------+--------------+--------------------+--------------+
2999 | I | (i, i, i, i) | (i, i, i, i) | N/A |
3000 +--------------------+--------------+--------------------+--------------+
3001 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
3002 | | | [#envmap-bumpmap]_ | |
3003 +--------------------+--------------+--------------------+--------------+
3004 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
3005 | | | [#depth-tex-mode]_ | |
3006 +--------------------+--------------+--------------------+--------------+
3007 | S | (s, s, s, s) | unknown | unknown |
3008 +--------------------+--------------+--------------------+--------------+
3010 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
3011 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
3012 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.