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:: RFL - Reflection Vector
483 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
485 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
487 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
493 Considered for removal.
496 .. opcode:: SEQ - Set On Equal
500 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
502 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
504 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
506 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
509 .. opcode:: SFL - Set On False
511 This instruction replicates its result.
519 Considered for removal.
522 .. opcode:: SGT - Set On Greater Than
526 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
528 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
530 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
532 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
535 .. opcode:: SIN - Sine
537 This instruction replicates its result.
544 .. opcode:: SLE - Set On Less Equal Than
548 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
550 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
552 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
554 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
557 .. opcode:: SNE - Set On Not Equal
561 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
563 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
565 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
567 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
570 .. opcode:: STR - Set On True
572 This instruction replicates its result.
579 .. opcode:: TEX - Texture Lookup
581 for array textures src0.y contains the slice for 1D,
582 and src0.z contain the slice for 2D.
584 for shadow textures with no arrays (and not cube map),
585 src0.z contains the reference value.
587 for shadow textures with arrays, src0.z contains
588 the reference value for 1D arrays, and src0.w contains
589 the reference value for 2D arrays and cube maps.
591 for cube map array shadow textures, the reference value
592 cannot be passed in src0.w, and TEX2 must be used instead.
598 shadow_ref = src0.z or src0.w (optional)
602 dst = texture\_sample(unit, coord, shadow_ref)
605 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
607 this is the same as TEX, but uses another reg to encode the
618 dst = texture\_sample(unit, coord, shadow_ref)
623 .. opcode:: TXD - Texture Lookup with Derivatives
635 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
638 .. opcode:: TXP - Projective Texture Lookup
642 coord.x = src0.x / src0.w
644 coord.y = src0.y / src0.w
646 coord.z = src0.z / src0.w
652 dst = texture\_sample(unit, coord)
655 .. opcode:: UP2H - Unpack Two 16-Bit Floats
661 Considered for removal.
663 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
669 Considered for removal.
671 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
677 Considered for removal.
679 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
685 Considered for removal.
687 .. opcode:: X2D - 2D Coordinate Transformation
691 dst.x = src0.x + src1.x \times src2.x + src1.y \times src2.y
693 dst.y = src0.y + src1.x \times src2.z + src1.y \times src2.w
695 dst.z = src0.x + src1.x \times src2.x + src1.y \times src2.y
697 dst.w = src0.y + src1.x \times src2.z + src1.y \times src2.w
701 Considered for removal.
704 .. opcode:: ARA - Address Register Add
710 Considered for removal.
712 .. opcode:: ARR - Address Register Load With Round
725 .. opcode:: SSG - Set Sign
729 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
731 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
733 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
735 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
738 .. opcode:: CMP - Compare
742 dst.x = (src0.x < 0) ? src1.x : src2.x
744 dst.y = (src0.y < 0) ? src1.y : src2.y
746 dst.z = (src0.z < 0) ? src1.z : src2.z
748 dst.w = (src0.w < 0) ? src1.w : src2.w
751 .. opcode:: KILL_IF - Conditional Discard
753 Conditional discard. Allowed in fragment shaders only.
757 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
762 .. opcode:: KILL - Discard
764 Unconditional discard. Allowed in fragment shaders only.
767 .. opcode:: SCS - Sine Cosine
780 .. opcode:: TXB - Texture Lookup With Bias
782 for cube map array textures and shadow cube maps, the bias value
783 cannot be passed in src0.w, and TXB2 must be used instead.
785 if the target is a shadow texture, the reference value is always
786 in src.z (this prevents shadow 3d and shadow 2d arrays from
787 using this instruction, but this is not needed).
803 dst = texture\_sample(unit, coord, bias)
806 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
808 this is the same as TXB, but uses another reg to encode the
809 lod bias value for cube map arrays and shadow cube maps.
810 Presumably shadow 2d arrays and shadow 3d targets could use
811 this encoding too, but this is not legal.
813 shadow cube map arrays are neither possible nor required.
823 dst = texture\_sample(unit, coord, bias)
826 .. opcode:: DIV - Divide
830 dst.x = \frac{src0.x}{src1.x}
832 dst.y = \frac{src0.y}{src1.y}
834 dst.z = \frac{src0.z}{src1.z}
836 dst.w = \frac{src0.w}{src1.w}
839 .. opcode:: DP2 - 2-component Dot Product
841 This instruction replicates its result.
845 dst = src0.x \times src1.x + src0.y \times src1.y
848 .. opcode:: TXL - Texture Lookup With explicit LOD
850 for cube map array textures, the explicit lod value
851 cannot be passed in src0.w, and TXL2 must be used instead.
853 if the target is a shadow texture, the reference value is always
854 in src.z (this prevents shadow 3d / 2d array / cube targets from
855 using this instruction, but this is not needed).
871 dst = texture\_sample(unit, coord, lod)
874 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
876 this is the same as TXL, but uses another reg to encode the
878 Presumably shadow 3d / 2d array / cube targets could use
879 this encoding too, but this is not legal.
881 shadow cube map arrays are neither possible nor required.
891 dst = texture\_sample(unit, coord, lod)
894 .. opcode:: PUSHA - Push Address Register On Stack
903 Considered for cleanup.
907 Considered for removal.
909 .. opcode:: POPA - Pop Address Register From Stack
918 Considered for cleanup.
922 Considered for removal.
925 .. opcode:: BRA - Branch
931 Considered for removal.
934 .. opcode:: CALLNZ - Subroutine Call If Not Zero
940 Considered for cleanup.
944 Considered for removal.
948 ^^^^^^^^^^^^^^^^^^^^^^^^
950 These opcodes are primarily provided for special-use computational shaders.
951 Support for these opcodes indicated by a special pipe capability bit (TBD).
953 XXX doesn't look like most of the opcodes really belong here.
955 .. opcode:: CEIL - Ceiling
959 dst.x = \lceil src.x\rceil
961 dst.y = \lceil src.y\rceil
963 dst.z = \lceil src.z\rceil
965 dst.w = \lceil src.w\rceil
968 .. opcode:: TRUNC - Truncate
981 .. opcode:: MOD - Modulus
985 dst.x = src0.x \bmod src1.x
987 dst.y = src0.y \bmod src1.y
989 dst.z = src0.z \bmod src1.z
991 dst.w = src0.w \bmod src1.w
994 .. opcode:: UARL - Integer Address Register Load
996 Moves the contents of the source register, assumed to be an integer, into the
997 destination register, which is assumed to be an address (ADDR) register.
1000 .. opcode:: SAD - Sum Of Absolute Differences
1004 dst.x = |src0.x - src1.x| + src2.x
1006 dst.y = |src0.y - src1.y| + src2.y
1008 dst.z = |src0.z - src1.z| + src2.z
1010 dst.w = |src0.w - src1.w| + src2.w
1013 .. opcode:: TXF - Texel Fetch
1015 As per NV_gpu_shader4, extract a single texel from a specified texture
1016 image. The source sampler may not be a CUBE or SHADOW. src 0 is a
1017 four-component signed integer vector used to identify the single texel
1018 accessed. 3 components + level. Just like texture instructions, an optional
1019 offset vector is provided, which is subject to various driver restrictions
1020 (regarding range, source of offsets).
1021 TXF(uint_vec coord, int_vec offset).
1024 .. opcode:: TXQ - Texture Size Query
1026 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
1027 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
1028 depth), 1D array (width, layers), 2D array (width, height, layers).
1029 Also return the number of accessible levels (last_level - first_level + 1)
1032 For components which don't return a resource dimension, their value
1040 dst.x = texture\_width(unit, lod)
1042 dst.y = texture\_height(unit, lod)
1044 dst.z = texture\_depth(unit, lod)
1046 dst.w = texture\_levels(unit)
1048 .. opcode:: TG4 - Texture Gather
1050 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
1051 filtering operation and packs them into a single register. Only works with
1052 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
1053 addressing modes of the sampler and the top level of any mip pyramid are
1054 used. Set W to zero. It behaves like the TEX instruction, but a filtered
1055 sample is not generated. The four samples that contribute to filtering are
1056 placed into xyzw in clockwise order, starting with the (u,v) texture
1057 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
1058 where the magnitude of the deltas are half a texel.
1060 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
1061 depth compares, single component selection, and a non-constant offset. It
1062 doesn't allow support for the GL independent offset to get i0,j0. This would
1063 require another CAP is hw can do it natively. For now we lower that before
1072 dst = texture\_gather4 (unit, coord, component)
1074 (with SM5 - cube array shadow)
1082 dst = texture\_gather (uint, coord, compare)
1084 .. opcode:: LODQ - level of detail query
1086 Compute the LOD information that the texture pipe would use to access the
1087 texture. The Y component contains the computed LOD lambda_prime. The X
1088 component contains the LOD that will be accessed, based on min/max lod's
1095 dst.xy = lodq(uint, coord);
1098 ^^^^^^^^^^^^^^^^^^^^^^^^
1099 These opcodes are used for integer operations.
1100 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1103 .. opcode:: I2F - Signed Integer To Float
1105 Rounding is unspecified (round to nearest even suggested).
1109 dst.x = (float) src.x
1111 dst.y = (float) src.y
1113 dst.z = (float) src.z
1115 dst.w = (float) src.w
1118 .. opcode:: U2F - Unsigned Integer To Float
1120 Rounding is unspecified (round to nearest even suggested).
1124 dst.x = (float) src.x
1126 dst.y = (float) src.y
1128 dst.z = (float) src.z
1130 dst.w = (float) src.w
1133 .. opcode:: F2I - Float to Signed Integer
1135 Rounding is towards zero (truncate).
1136 Values outside signed range (including NaNs) produce undefined results.
1149 .. opcode:: F2U - Float to Unsigned Integer
1151 Rounding is towards zero (truncate).
1152 Values outside unsigned range (including NaNs) produce undefined results.
1156 dst.x = (unsigned) src.x
1158 dst.y = (unsigned) src.y
1160 dst.z = (unsigned) src.z
1162 dst.w = (unsigned) src.w
1165 .. opcode:: UADD - Integer Add
1167 This instruction works the same for signed and unsigned integers.
1168 The low 32bit of the result is returned.
1172 dst.x = src0.x + src1.x
1174 dst.y = src0.y + src1.y
1176 dst.z = src0.z + src1.z
1178 dst.w = src0.w + src1.w
1181 .. opcode:: UMAD - Integer Multiply And Add
1183 This instruction works the same for signed and unsigned integers.
1184 The multiplication returns the low 32bit (as does the result itself).
1188 dst.x = src0.x \times src1.x + src2.x
1190 dst.y = src0.y \times src1.y + src2.y
1192 dst.z = src0.z \times src1.z + src2.z
1194 dst.w = src0.w \times src1.w + src2.w
1197 .. opcode:: UMUL - Integer Multiply
1199 This instruction works the same for signed and unsigned integers.
1200 The low 32bit of the result is returned.
1204 dst.x = src0.x \times src1.x
1206 dst.y = src0.y \times src1.y
1208 dst.z = src0.z \times src1.z
1210 dst.w = src0.w \times src1.w
1213 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1215 The high 32bits of the multiplication of 2 signed integers are returned.
1219 dst.x = (src0.x \times src1.x) >> 32
1221 dst.y = (src0.y \times src1.y) >> 32
1223 dst.z = (src0.z \times src1.z) >> 32
1225 dst.w = (src0.w \times src1.w) >> 32
1228 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1230 The high 32bits of the multiplication of 2 unsigned integers are returned.
1234 dst.x = (src0.x \times src1.x) >> 32
1236 dst.y = (src0.y \times src1.y) >> 32
1238 dst.z = (src0.z \times src1.z) >> 32
1240 dst.w = (src0.w \times src1.w) >> 32
1243 .. opcode:: IDIV - Signed Integer Division
1245 TBD: behavior for division by zero.
1249 dst.x = src0.x \ src1.x
1251 dst.y = src0.y \ src1.y
1253 dst.z = src0.z \ src1.z
1255 dst.w = src0.w \ src1.w
1258 .. opcode:: UDIV - Unsigned Integer Division
1260 For division by zero, 0xffffffff is returned.
1264 dst.x = src0.x \ src1.x
1266 dst.y = src0.y \ src1.y
1268 dst.z = src0.z \ src1.z
1270 dst.w = src0.w \ src1.w
1273 .. opcode:: UMOD - Unsigned Integer Remainder
1275 If second arg is zero, 0xffffffff is returned.
1279 dst.x = src0.x \ src1.x
1281 dst.y = src0.y \ src1.y
1283 dst.z = src0.z \ src1.z
1285 dst.w = src0.w \ src1.w
1288 .. opcode:: NOT - Bitwise Not
1301 .. opcode:: AND - Bitwise And
1305 dst.x = src0.x \& src1.x
1307 dst.y = src0.y \& src1.y
1309 dst.z = src0.z \& src1.z
1311 dst.w = src0.w \& src1.w
1314 .. opcode:: OR - Bitwise Or
1318 dst.x = src0.x | src1.x
1320 dst.y = src0.y | src1.y
1322 dst.z = src0.z | src1.z
1324 dst.w = src0.w | src1.w
1327 .. opcode:: XOR - Bitwise Xor
1331 dst.x = src0.x \oplus src1.x
1333 dst.y = src0.y \oplus src1.y
1335 dst.z = src0.z \oplus src1.z
1337 dst.w = src0.w \oplus src1.w
1340 .. opcode:: IMAX - Maximum of Signed Integers
1344 dst.x = max(src0.x, src1.x)
1346 dst.y = max(src0.y, src1.y)
1348 dst.z = max(src0.z, src1.z)
1350 dst.w = max(src0.w, src1.w)
1353 .. opcode:: UMAX - Maximum of Unsigned Integers
1357 dst.x = max(src0.x, src1.x)
1359 dst.y = max(src0.y, src1.y)
1361 dst.z = max(src0.z, src1.z)
1363 dst.w = max(src0.w, src1.w)
1366 .. opcode:: IMIN - Minimum of Signed Integers
1370 dst.x = min(src0.x, src1.x)
1372 dst.y = min(src0.y, src1.y)
1374 dst.z = min(src0.z, src1.z)
1376 dst.w = min(src0.w, src1.w)
1379 .. opcode:: UMIN - Minimum of Unsigned Integers
1383 dst.x = min(src0.x, src1.x)
1385 dst.y = min(src0.y, src1.y)
1387 dst.z = min(src0.z, src1.z)
1389 dst.w = min(src0.w, src1.w)
1392 .. opcode:: SHL - Shift Left
1394 The shift count is masked with 0x1f before the shift is applied.
1398 dst.x = src0.x << (0x1f \& src1.x)
1400 dst.y = src0.y << (0x1f \& src1.y)
1402 dst.z = src0.z << (0x1f \& src1.z)
1404 dst.w = src0.w << (0x1f \& src1.w)
1407 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1409 The shift count is masked with 0x1f before the shift is applied.
1413 dst.x = src0.x >> (0x1f \& src1.x)
1415 dst.y = src0.y >> (0x1f \& src1.y)
1417 dst.z = src0.z >> (0x1f \& src1.z)
1419 dst.w = src0.w >> (0x1f \& src1.w)
1422 .. opcode:: USHR - Logical Shift Right
1424 The shift count is masked with 0x1f before the shift is applied.
1428 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1430 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1432 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1434 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1437 .. opcode:: UCMP - Integer Conditional Move
1441 dst.x = src0.x ? src1.x : src2.x
1443 dst.y = src0.y ? src1.y : src2.y
1445 dst.z = src0.z ? src1.z : src2.z
1447 dst.w = src0.w ? src1.w : src2.w
1451 .. opcode:: ISSG - Integer Set Sign
1455 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1457 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1459 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1461 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1465 .. opcode:: FSLT - Float Set On Less Than (ordered)
1467 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1471 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1473 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1475 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1477 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1480 .. opcode:: ISLT - Signed Integer Set On Less Than
1484 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1486 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1488 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1490 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1493 .. opcode:: USLT - Unsigned Integer Set On Less Than
1497 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1499 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1501 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1503 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1506 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1508 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1512 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1514 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1516 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1518 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1521 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1525 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1527 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1529 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1531 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1534 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1538 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1540 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1542 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1544 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1547 .. opcode:: FSEQ - Float Set On Equal (ordered)
1549 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1553 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1555 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1557 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1559 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1562 .. opcode:: USEQ - Integer Set On Equal
1566 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1568 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1570 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1572 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1575 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1577 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1581 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1583 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1585 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1587 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1590 .. opcode:: USNE - Integer Set On Not Equal
1594 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1596 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1598 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1600 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1603 .. opcode:: INEG - Integer Negate
1618 .. opcode:: IABS - Integer Absolute Value
1632 These opcodes are used for bit-level manipulation of integers.
1634 .. opcode:: IBFE - Signed Bitfield Extract
1636 See SM5 instruction of the same name. Extracts a set of bits from the input,
1637 and sign-extends them if the high bit of the extracted window is set.
1641 def ibfe(value, offset, bits):
1642 offset = offset & 0x1f
1644 if bits == 0: return 0
1645 # Note: >> sign-extends
1646 if width + offset < 32:
1647 return (value << (32 - offset - bits)) >> (32 - bits)
1649 return value >> offset
1651 .. opcode:: UBFE - Unsigned Bitfield Extract
1653 See SM5 instruction of the same name. Extracts a set of bits from the input,
1654 without any sign-extension.
1658 def ubfe(value, offset, bits):
1659 offset = offset & 0x1f
1661 if bits == 0: return 0
1662 # Note: >> does not sign-extend
1663 if width + offset < 32:
1664 return (value << (32 - offset - bits)) >> (32 - bits)
1666 return value >> offset
1668 .. opcode:: BFI - Bitfield Insert
1670 See SM5 instruction of the same name. Replaces a bit region of 'base' with
1671 the low bits of 'insert'.
1675 def bfi(base, insert, offset, bits):
1676 offset = offset & 0x1f
1678 mask = ((1 << bits) - 1) << offset
1679 return ((insert << offset) & mask) | (base & ~mask)
1681 .. opcode:: BREV - Bitfield Reverse
1683 See SM5 instruction BFREV. Reverses the bits of the argument.
1685 .. opcode:: POPC - Population Count
1687 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1689 .. opcode:: LSB - Index of lowest set bit
1691 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1692 bit of the argument. Returns -1 if none are set.
1694 .. opcode:: IMSB - Index of highest non-sign bit
1696 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1697 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1698 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1699 (i.e. for inputs 0 and -1).
1701 .. opcode:: UMSB - Index of highest set bit
1703 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1704 set bit of the argument. Returns -1 if none are set.
1707 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1709 These opcodes are only supported in geometry shaders; they have no meaning
1710 in any other type of shader.
1712 .. opcode:: EMIT - Emit
1714 Generate a new vertex for the current primitive into the specified vertex
1715 stream using the values in the output registers.
1718 .. opcode:: ENDPRIM - End Primitive
1720 Complete the current primitive in the specified vertex stream (consisting of
1721 the emitted vertices), and start a new one.
1727 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1728 opcodes is determined by a special capability bit, ``GLSL``.
1729 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1731 .. opcode:: CAL - Subroutine Call
1737 .. opcode:: RET - Subroutine Call Return
1742 .. opcode:: CONT - Continue
1744 Unconditionally moves the point of execution to the instruction after the
1745 last bgnloop. The instruction must appear within a bgnloop/endloop.
1749 Support for CONT is determined by a special capability bit,
1750 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1753 .. opcode:: BGNLOOP - Begin a Loop
1755 Start a loop. Must have a matching endloop.
1758 .. opcode:: BGNSUB - Begin Subroutine
1760 Starts definition of a subroutine. Must have a matching endsub.
1763 .. opcode:: ENDLOOP - End a Loop
1765 End a loop started with bgnloop.
1768 .. opcode:: ENDSUB - End Subroutine
1770 Ends definition of a subroutine.
1773 .. opcode:: NOP - No Operation
1778 .. opcode:: BRK - Break
1780 Unconditionally moves the point of execution to the instruction after the
1781 next endloop or endswitch. The instruction must appear within a loop/endloop
1782 or switch/endswitch.
1785 .. opcode:: BREAKC - Break Conditional
1787 Conditionally moves the point of execution to the instruction after the
1788 next endloop or endswitch. The instruction must appear within a loop/endloop
1789 or switch/endswitch.
1790 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1791 as an integer register.
1795 Considered for removal as it's quite inconsistent wrt other opcodes
1796 (could emulate with UIF/BRK/ENDIF).
1799 .. opcode:: IF - Float If
1801 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1805 where src0.x is interpreted as a floating point register.
1808 .. opcode:: UIF - Bitwise If
1810 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1814 where src0.x is interpreted as an integer register.
1817 .. opcode:: ELSE - Else
1819 Starts an else block, after an IF or UIF statement.
1822 .. opcode:: ENDIF - End If
1824 Ends an IF or UIF block.
1827 .. opcode:: SWITCH - Switch
1829 Starts a C-style switch expression. The switch consists of one or multiple
1830 CASE statements, and at most one DEFAULT statement. Execution of a statement
1831 ends when a BRK is hit, but just like in C falling through to other cases
1832 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1833 just as last statement, and fallthrough is allowed into/from it.
1834 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1840 (some instructions here)
1843 (some instructions here)
1846 (some instructions here)
1851 .. opcode:: CASE - Switch case
1853 This represents a switch case label. The src arg must be an integer immediate.
1856 .. opcode:: DEFAULT - Switch default
1858 This represents the default case in the switch, which is taken if no other
1862 .. opcode:: ENDSWITCH - End of switch
1864 Ends a switch expression.
1870 The interpolation instructions allow an input to be interpolated in a
1871 different way than its declaration. This corresponds to the GLSL 4.00
1872 interpolateAt* functions. The first argument of each of these must come from
1873 ``TGSI_FILE_INPUT``.
1875 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1877 Interpolates the varying specified by src0 at the centroid
1879 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1881 Interpolates the varying specified by src0 at the sample id specified by
1882 src1.x (interpreted as an integer)
1884 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1886 Interpolates the varying specified by src0 at the offset src1.xy from the
1887 pixel center (interpreted as floats)
1895 The double-precision opcodes reinterpret four-component vectors into
1896 two-component vectors with doubled precision in each component.
1898 Support for these opcodes is XXX undecided. :T
1900 .. opcode:: DADD - Add
1904 dst.xy = src0.xy + src1.xy
1906 dst.zw = src0.zw + src1.zw
1909 .. opcode:: DDIV - Divide
1913 dst.xy = src0.xy / src1.xy
1915 dst.zw = src0.zw / src1.zw
1917 .. opcode:: DSEQ - Set on Equal
1921 dst.xy = src0.xy == src1.xy ? 1.0F : 0.0F
1923 dst.zw = src0.zw == src1.zw ? 1.0F : 0.0F
1925 .. opcode:: DSLT - Set on Less than
1929 dst.xy = src0.xy < src1.xy ? 1.0F : 0.0F
1931 dst.zw = src0.zw < src1.zw ? 1.0F : 0.0F
1933 .. opcode:: DFRAC - Fraction
1937 dst.xy = src.xy - \lfloor src.xy\rfloor
1939 dst.zw = src.zw - \lfloor src.zw\rfloor
1942 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1944 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1945 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1946 :math:`dst1 \times 2^{dst0} = src` .
1950 dst0.xy = exp(src.xy)
1952 dst1.xy = frac(src.xy)
1954 dst0.zw = exp(src.zw)
1956 dst1.zw = frac(src.zw)
1958 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1960 This opcode is the inverse of :opcode:`DFRACEXP`.
1964 dst.xy = src0.xy \times 2^{src1.xy}
1966 dst.zw = src0.zw \times 2^{src1.zw}
1968 .. opcode:: DMIN - Minimum
1972 dst.xy = min(src0.xy, src1.xy)
1974 dst.zw = min(src0.zw, src1.zw)
1976 .. opcode:: DMAX - Maximum
1980 dst.xy = max(src0.xy, src1.xy)
1982 dst.zw = max(src0.zw, src1.zw)
1984 .. opcode:: DMUL - Multiply
1988 dst.xy = src0.xy \times src1.xy
1990 dst.zw = src0.zw \times src1.zw
1993 .. opcode:: DMAD - Multiply And Add
1997 dst.xy = src0.xy \times src1.xy + src2.xy
1999 dst.zw = src0.zw \times src1.zw + src2.zw
2002 .. opcode:: DRCP - Reciprocal
2006 dst.xy = \frac{1}{src.xy}
2008 dst.zw = \frac{1}{src.zw}
2010 .. opcode:: DSQRT - Square Root
2014 dst.xy = \sqrt{src.xy}
2016 dst.zw = \sqrt{src.zw}
2019 .. _samplingopcodes:
2021 Resource Sampling Opcodes
2022 ^^^^^^^^^^^^^^^^^^^^^^^^^
2024 Those opcodes follow very closely semantics of the respective Direct3D
2025 instructions. If in doubt double check Direct3D documentation.
2026 Note that the swizzle on SVIEW (src1) determines texel swizzling
2031 Using provided address, sample data from the specified texture using the
2032 filtering mode identified by the gven sampler. The source data may come from
2033 any resource type other than buffers.
2035 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
2037 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
2039 .. opcode:: SAMPLE_I
2041 Simplified alternative to the SAMPLE instruction. Using the provided
2042 integer address, SAMPLE_I fetches data from the specified sampler view
2043 without any filtering. The source data may come from any resource type
2046 Syntax: ``SAMPLE_I dst, address, sampler_view``
2048 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
2050 The 'address' is specified as unsigned integers. If the 'address' is out of
2051 range [0...(# texels - 1)] the result of the fetch is always 0 in all
2052 components. As such the instruction doesn't honor address wrap modes, in
2053 cases where that behavior is desirable 'SAMPLE' instruction should be used.
2054 address.w always provides an unsigned integer mipmap level. If the value is
2055 out of the range then the instruction always returns 0 in all components.
2056 address.yz are ignored for buffers and 1d textures. address.z is ignored
2057 for 1d texture arrays and 2d textures.
2059 For 1D texture arrays address.y provides the array index (also as unsigned
2060 integer). If the value is out of the range of available array indices
2061 [0... (array size - 1)] then the opcode always returns 0 in all components.
2062 For 2D texture arrays address.z provides the array index, otherwise it
2063 exhibits the same behavior as in the case for 1D texture arrays. The exact
2064 semantics of the source address are presented in the table below:
2066 +---------------------------+----+-----+-----+---------+
2067 | resource type | X | Y | Z | W |
2068 +===========================+====+=====+=====+=========+
2069 | ``PIPE_BUFFER`` | x | | | ignored |
2070 +---------------------------+----+-----+-----+---------+
2071 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2072 +---------------------------+----+-----+-----+---------+
2073 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2074 +---------------------------+----+-----+-----+---------+
2075 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2076 +---------------------------+----+-----+-----+---------+
2077 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2078 +---------------------------+----+-----+-----+---------+
2079 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2080 +---------------------------+----+-----+-----+---------+
2081 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2082 +---------------------------+----+-----+-----+---------+
2083 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2084 +---------------------------+----+-----+-----+---------+
2086 Where 'mpl' is a mipmap level and 'idx' is the array index.
2088 .. opcode:: SAMPLE_I_MS
2090 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2092 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2094 .. opcode:: SAMPLE_B
2096 Just like the SAMPLE instruction with the exception that an additional bias
2097 is applied to the level of detail computed as part of the instruction
2100 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2102 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2104 .. opcode:: SAMPLE_C
2106 Similar to the SAMPLE instruction but it performs a comparison filter. The
2107 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2108 additional float32 operand, reference value, which must be a register with
2109 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2110 current samplers compare_func (in pipe_sampler_state) to compare reference
2111 value against the red component value for the surce resource at each texel
2112 that the currently configured texture filter covers based on the provided
2115 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2117 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2119 .. opcode:: SAMPLE_C_LZ
2121 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2124 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2126 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2129 .. opcode:: SAMPLE_D
2131 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2132 the source address in the x direction and the y direction are provided by
2135 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2137 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2139 .. opcode:: SAMPLE_L
2141 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2142 directly as a scalar value, representing no anisotropy.
2144 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2146 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2150 Gathers the four texels to be used in a bi-linear filtering operation and
2151 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2152 and cubemaps arrays. For 2D textures, only the addressing modes of the
2153 sampler and the top level of any mip pyramid are used. Set W to zero. It
2154 behaves like the SAMPLE instruction, but a filtered sample is not
2155 generated. The four samples that contribute to filtering are placed into
2156 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2157 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2158 magnitude of the deltas are half a texel.
2161 .. opcode:: SVIEWINFO
2163 Query the dimensions of a given sampler view. dst receives width, height,
2164 depth or array size and number of mipmap levels as int4. The dst can have a
2165 writemask which will specify what info is the caller interested in.
2167 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2169 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2171 src_mip_level is an unsigned integer scalar. If it's out of range then
2172 returns 0 for width, height and depth/array size but the total number of
2173 mipmap is still returned correctly for the given sampler view. The returned
2174 width, height and depth values are for the mipmap level selected by the
2175 src_mip_level and are in the number of texels. For 1d texture array width
2176 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2177 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2178 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2179 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2180 resinfo allowing swizzling dst values is ignored (due to the interaction
2181 with rcpfloat modifier which requires some swizzle handling in the state
2184 .. opcode:: SAMPLE_POS
2186 Query the position of a given sample. dst receives float4 (x, y, 0, 0)
2187 indicated where the sample is located. If the resource is not a multi-sample
2188 resource and not a render target, the result is 0.
2190 .. opcode:: SAMPLE_INFO
2192 dst receives number of samples in x. If the resource is not a multi-sample
2193 resource and not a render target, the result is 0.
2196 .. _resourceopcodes:
2198 Resource Access Opcodes
2199 ^^^^^^^^^^^^^^^^^^^^^^^
2201 .. opcode:: LOAD - Fetch data from a shader resource
2203 Syntax: ``LOAD dst, resource, address``
2205 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
2207 Using the provided integer address, LOAD fetches data
2208 from the specified buffer or texture without any
2211 The 'address' is specified as a vector of unsigned
2212 integers. If the 'address' is out of range the result
2215 Only the first mipmap level of a resource can be read
2216 from using this instruction.
2218 For 1D or 2D texture arrays, the array index is
2219 provided as an unsigned integer in address.y or
2220 address.z, respectively. address.yz are ignored for
2221 buffers and 1D textures. address.z is ignored for 1D
2222 texture arrays and 2D textures. address.w is always
2225 .. opcode:: STORE - Write data to a shader resource
2227 Syntax: ``STORE resource, address, src``
2229 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2231 Using the provided integer address, STORE writes data
2232 to the specified buffer or texture.
2234 The 'address' is specified as a vector of unsigned
2235 integers. If the 'address' is out of range the result
2238 Only the first mipmap level of a resource can be
2239 written to using this instruction.
2241 For 1D or 2D texture arrays, the array index is
2242 provided as an unsigned integer in address.y or
2243 address.z, respectively. address.yz are ignored for
2244 buffers and 1D textures. address.z is ignored for 1D
2245 texture arrays and 2D textures. address.w is always
2249 .. _threadsyncopcodes:
2251 Inter-thread synchronization opcodes
2252 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2254 These opcodes are intended for communication between threads running
2255 within the same compute grid. For now they're only valid in compute
2258 .. opcode:: MFENCE - Memory fence
2260 Syntax: ``MFENCE resource``
2262 Example: ``MFENCE RES[0]``
2264 This opcode forces strong ordering between any memory access
2265 operations that affect the specified resource. This means that
2266 previous loads and stores (and only those) will be performed and
2267 visible to other threads before the program execution continues.
2270 .. opcode:: LFENCE - Load memory fence
2272 Syntax: ``LFENCE resource``
2274 Example: ``LFENCE RES[0]``
2276 Similar to MFENCE, but it only affects the ordering of memory loads.
2279 .. opcode:: SFENCE - Store memory fence
2281 Syntax: ``SFENCE resource``
2283 Example: ``SFENCE RES[0]``
2285 Similar to MFENCE, but it only affects the ordering of memory stores.
2288 .. opcode:: BARRIER - Thread group barrier
2292 This opcode suspends the execution of the current thread until all
2293 the remaining threads in the working group reach the same point of
2294 the program. Results are unspecified if any of the remaining
2295 threads terminates or never reaches an executed BARRIER instruction.
2303 These opcodes provide atomic variants of some common arithmetic and
2304 logical operations. In this context atomicity means that another
2305 concurrent memory access operation that affects the same memory
2306 location is guaranteed to be performed strictly before or after the
2307 entire execution of the atomic operation.
2309 For the moment they're only valid in compute programs.
2311 .. opcode:: ATOMUADD - Atomic integer addition
2313 Syntax: ``ATOMUADD dst, resource, offset, src``
2315 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2317 The following operation is performed atomically on each component:
2321 dst_i = resource[offset]_i
2323 resource[offset]_i = dst_i + src_i
2326 .. opcode:: ATOMXCHG - Atomic exchange
2328 Syntax: ``ATOMXCHG dst, resource, offset, src``
2330 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2332 The following operation is performed atomically on each component:
2336 dst_i = resource[offset]_i
2338 resource[offset]_i = src_i
2341 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2343 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2345 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2347 The following operation is performed atomically on each component:
2351 dst_i = resource[offset]_i
2353 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2356 .. opcode:: ATOMAND - Atomic bitwise And
2358 Syntax: ``ATOMAND dst, resource, offset, src``
2360 Example: ``ATOMAND TEMP[0], RES[0], TEMP[1], TEMP[2]``
2362 The following operation is performed atomically on each component:
2366 dst_i = resource[offset]_i
2368 resource[offset]_i = dst_i \& src_i
2371 .. opcode:: ATOMOR - Atomic bitwise Or
2373 Syntax: ``ATOMOR dst, resource, offset, src``
2375 Example: ``ATOMOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2377 The following operation is performed atomically on each component:
2381 dst_i = resource[offset]_i
2383 resource[offset]_i = dst_i | src_i
2386 .. opcode:: ATOMXOR - Atomic bitwise Xor
2388 Syntax: ``ATOMXOR dst, resource, offset, src``
2390 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2392 The following operation is performed atomically on each component:
2396 dst_i = resource[offset]_i
2398 resource[offset]_i = dst_i \oplus src_i
2401 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2403 Syntax: ``ATOMUMIN dst, resource, offset, src``
2405 Example: ``ATOMUMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2407 The following operation is performed atomically on each component:
2411 dst_i = resource[offset]_i
2413 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2416 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2418 Syntax: ``ATOMUMAX dst, resource, offset, src``
2420 Example: ``ATOMUMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2422 The following operation is performed atomically on each component:
2426 dst_i = resource[offset]_i
2428 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2431 .. opcode:: ATOMIMIN - Atomic signed minimum
2433 Syntax: ``ATOMIMIN dst, resource, offset, src``
2435 Example: ``ATOMIMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2437 The following operation is performed atomically on each component:
2441 dst_i = resource[offset]_i
2443 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2446 .. opcode:: ATOMIMAX - Atomic signed maximum
2448 Syntax: ``ATOMIMAX dst, resource, offset, src``
2450 Example: ``ATOMIMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2452 The following operation is performed atomically on each component:
2456 dst_i = resource[offset]_i
2458 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2462 Explanation of symbols used
2463 ------------------------------
2470 :math:`|x|` Absolute value of `x`.
2472 :math:`\lceil x \rceil` Ceiling of `x`.
2474 clamp(x,y,z) Clamp x between y and z.
2475 (x < y) ? y : (x > z) ? z : x
2477 :math:`\lfloor x\rfloor` Floor of `x`.
2479 :math:`\log_2{x}` Logarithm of `x`, base 2.
2481 max(x,y) Maximum of x and y.
2484 min(x,y) Minimum of x and y.
2487 partialx(x) Derivative of x relative to fragment's X.
2489 partialy(x) Derivative of x relative to fragment's Y.
2491 pop() Pop from stack.
2493 :math:`x^y` `x` to the power `y`.
2495 push(x) Push x on stack.
2499 trunc(x) Truncate x, i.e. drop the fraction bits.
2506 discard Discard fragment.
2510 target Label of target instruction.
2521 Declares a register that is will be referenced as an operand in Instruction
2524 File field contains register file that is being declared and is one
2527 UsageMask field specifies which of the register components can be accessed
2528 and is one of TGSI_WRITEMASK.
2530 The Local flag specifies that a given value isn't intended for
2531 subroutine parameter passing and, as a result, the implementation
2532 isn't required to give any guarantees of it being preserved across
2533 subroutine boundaries. As it's merely a compiler hint, the
2534 implementation is free to ignore it.
2536 If Dimension flag is set to 1, a Declaration Dimension token follows.
2538 If Semantic flag is set to 1, a Declaration Semantic token follows.
2540 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2542 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2544 If Array flag is set to 1, a Declaration Array token follows.
2547 ^^^^^^^^^^^^^^^^^^^^^^^^
2549 Declarations can optional have an ArrayID attribute which can be referred by
2550 indirect addressing operands. An ArrayID of zero is reserved and treaded as
2551 if no ArrayID is specified.
2553 If an indirect addressing operand refers to a specific declaration by using
2554 an ArrayID only the registers in this declaration are guaranteed to be
2555 accessed, accessing any register outside this declaration results in undefined
2556 behavior. Note that for compatibility the effective index is zero-based and
2557 not relative to the specified declaration
2559 If no ArrayID is specified with an indirect addressing operand the whole
2560 register file might be accessed by this operand. This is strongly discouraged
2561 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2563 Declaration Semantic
2564 ^^^^^^^^^^^^^^^^^^^^^^^^
2566 Vertex and fragment shader input and output registers may be labeled
2567 with semantic information consisting of a name and index.
2569 Follows Declaration token if Semantic bit is set.
2571 Since its purpose is to link a shader with other stages of the pipeline,
2572 it is valid to follow only those Declaration tokens that declare a register
2573 either in INPUT or OUTPUT file.
2575 SemanticName field contains the semantic name of the register being declared.
2576 There is no default value.
2578 SemanticIndex is an optional subscript that can be used to distinguish
2579 different register declarations with the same semantic name. The default value
2582 The meanings of the individual semantic names are explained in the following
2585 TGSI_SEMANTIC_POSITION
2586 """"""""""""""""""""""
2588 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2589 output register which contains the homogeneous vertex position in the clip
2590 space coordinate system. After clipping, the X, Y and Z components of the
2591 vertex will be divided by the W value to get normalized device coordinates.
2593 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2594 fragment shader input contains the fragment's window position. The X
2595 component starts at zero and always increases from left to right.
2596 The Y component starts at zero and always increases but Y=0 may either
2597 indicate the top of the window or the bottom depending on the fragment
2598 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2599 The Z coordinate ranges from 0 to 1 to represent depth from the front
2600 to the back of the Z buffer. The W component contains the reciprocol
2601 of the interpolated vertex position W component.
2603 Fragment shaders may also declare an output register with
2604 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2605 the fragment shader to change the fragment's Z position.
2612 For vertex shader outputs or fragment shader inputs/outputs, this
2613 label indicates that the resister contains an R,G,B,A color.
2615 Several shader inputs/outputs may contain colors so the semantic index
2616 is used to distinguish them. For example, color[0] may be the diffuse
2617 color while color[1] may be the specular color.
2619 This label is needed so that the flat/smooth shading can be applied
2620 to the right interpolants during rasterization.
2624 TGSI_SEMANTIC_BCOLOR
2625 """"""""""""""""""""
2627 Back-facing colors are only used for back-facing polygons, and are only valid
2628 in vertex shader outputs. After rasterization, all polygons are front-facing
2629 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2630 so all BCOLORs effectively become regular COLORs in the fragment shader.
2636 Vertex shader inputs and outputs and fragment shader inputs may be
2637 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2638 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2639 to compute a fog blend factor which is used to blend the normal fragment color
2640 with a constant fog color. But fog coord really is just an ordinary vec4
2641 register like regular semantics.
2647 Vertex shader input and output registers may be labeled with
2648 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2649 in the form (S, 0, 0, 1). The point size controls the width or diameter
2650 of points for rasterization. This label cannot be used in fragment
2653 When using this semantic, be sure to set the appropriate state in the
2654 :ref:`rasterizer` first.
2657 TGSI_SEMANTIC_TEXCOORD
2658 """"""""""""""""""""""
2660 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2662 Vertex shader outputs and fragment shader inputs may be labeled with
2663 this semantic to make them replaceable by sprite coordinates via the
2664 sprite_coord_enable state in the :ref:`rasterizer`.
2665 The semantic index permitted with this semantic is limited to <= 7.
2667 If the driver does not support TEXCOORD, sprite coordinate replacement
2668 applies to inputs with the GENERIC semantic instead.
2670 The intended use case for this semantic is gl_TexCoord.
2673 TGSI_SEMANTIC_PCOORD
2674 """"""""""""""""""""
2676 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2678 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2679 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2680 the current primitive is a point and point sprites are enabled. Otherwise,
2681 the contents of the register are undefined.
2683 The intended use case for this semantic is gl_PointCoord.
2686 TGSI_SEMANTIC_GENERIC
2687 """""""""""""""""""""
2689 All vertex/fragment shader inputs/outputs not labeled with any other
2690 semantic label can be considered to be generic attributes. Typical
2691 uses of generic inputs/outputs are texcoords and user-defined values.
2694 TGSI_SEMANTIC_NORMAL
2695 """"""""""""""""""""
2697 Indicates that a vertex shader input is a normal vector. This is
2698 typically only used for legacy graphics APIs.
2704 This label applies to fragment shader inputs only and indicates that
2705 the register contains front/back-face information of the form (F, 0,
2706 0, 1). The first component will be positive when the fragment belongs
2707 to a front-facing polygon, and negative when the fragment belongs to a
2708 back-facing polygon.
2711 TGSI_SEMANTIC_EDGEFLAG
2712 """"""""""""""""""""""
2714 For vertex shaders, this sematic label indicates that an input or
2715 output is a boolean edge flag. The register layout is [F, x, x, x]
2716 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2717 simply copies the edge flag input to the edgeflag output.
2719 Edge flags are used to control which lines or points are actually
2720 drawn when the polygon mode converts triangles/quads/polygons into
2724 TGSI_SEMANTIC_STENCIL
2725 """""""""""""""""""""
2727 For fragment shaders, this semantic label indicates that an output
2728 is a writable stencil reference value. Only the Y component is writable.
2729 This allows the fragment shader to change the fragments stencilref value.
2732 TGSI_SEMANTIC_VIEWPORT_INDEX
2733 """"""""""""""""""""""""""""
2735 For geometry shaders, this semantic label indicates that an output
2736 contains the index of the viewport (and scissor) to use.
2737 Only the X value is used.
2743 For geometry shaders, this semantic label indicates that an output
2744 contains the layer value to use for the color and depth/stencil surfaces.
2745 Only the X value is used. (Also known as rendertarget array index.)
2748 TGSI_SEMANTIC_CULLDIST
2749 """"""""""""""""""""""
2751 Used as distance to plane for performing application-defined culling
2752 of individual primitives against a plane. When components of vertex
2753 elements are given this label, these values are assumed to be a
2754 float32 signed distance to a plane. Primitives will be completely
2755 discarded if the plane distance for all of the vertices in the
2756 primitive are < 0. If a vertex has a cull distance of NaN, that
2757 vertex counts as "out" (as if its < 0);
2758 The limits on both clip and cull distances are bound
2759 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2760 the maximum number of components that can be used to hold the
2761 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2762 which specifies the maximum number of registers which can be
2763 annotated with those semantics.
2766 TGSI_SEMANTIC_CLIPDIST
2767 """"""""""""""""""""""
2769 When components of vertex elements are identified this way, these
2770 values are each assumed to be a float32 signed distance to a plane.
2771 Primitive setup only invokes rasterization on pixels for which
2772 the interpolated plane distances are >= 0. Multiple clip planes
2773 can be implemented simultaneously, by annotating multiple
2774 components of one or more vertex elements with the above specified
2775 semantic. The limits on both clip and cull distances are bound
2776 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2777 the maximum number of components that can be used to hold the
2778 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2779 which specifies the maximum number of registers which can be
2780 annotated with those semantics.
2782 TGSI_SEMANTIC_SAMPLEID
2783 """"""""""""""""""""""
2785 For fragment shaders, this semantic label indicates that a system value
2786 contains the current sample id (i.e. gl_SampleID). Only the X value is used.
2788 TGSI_SEMANTIC_SAMPLEPOS
2789 """""""""""""""""""""""
2791 For fragment shaders, this semantic label indicates that a system value
2792 contains the current sample's position (i.e. gl_SamplePosition). Only the X
2793 and Y values are used.
2795 TGSI_SEMANTIC_SAMPLEMASK
2796 """"""""""""""""""""""""
2798 For fragment shaders, this semantic label indicates that an output contains
2799 the sample mask used to disable further sample processing
2800 (i.e. gl_SampleMask). Only the X value is used, up to 32x MS.
2802 TGSI_SEMANTIC_INVOCATIONID
2803 """"""""""""""""""""""""""
2805 For geometry shaders, this semantic label indicates that a system value
2806 contains the current invocation id (i.e. gl_InvocationID). Only the X value is
2809 Declaration Interpolate
2810 ^^^^^^^^^^^^^^^^^^^^^^^
2812 This token is only valid for fragment shader INPUT declarations.
2814 The Interpolate field specifes the way input is being interpolated by
2815 the rasteriser and is one of TGSI_INTERPOLATE_*.
2817 The Location field specifies the location inside the pixel that the
2818 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
2819 when per-sample shading is enabled, the implementation may choose to
2820 interpolate at the sample irrespective of the Location field.
2822 The CylindricalWrap bitfield specifies which register components
2823 should be subject to cylindrical wrapping when interpolating by the
2824 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2825 should be interpolated according to cylindrical wrapping rules.
2828 Declaration Sampler View
2829 ^^^^^^^^^^^^^^^^^^^^^^^^
2831 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2833 DCL SVIEW[#], resource, type(s)
2835 Declares a shader input sampler view and assigns it to a SVIEW[#]
2838 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2840 type must be 1 or 4 entries (if specifying on a per-component
2841 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
2844 Declaration Resource
2845 ^^^^^^^^^^^^^^^^^^^^
2847 Follows Declaration token if file is TGSI_FILE_RESOURCE.
2849 DCL RES[#], resource [, WR] [, RAW]
2851 Declares a shader input resource and assigns it to a RES[#]
2854 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
2857 If the RAW keyword is not specified, the texture data will be
2858 subject to conversion, swizzling and scaling as required to yield
2859 the specified data type from the physical data format of the bound
2862 If the RAW keyword is specified, no channel conversion will be
2863 performed: the values read for each of the channels (X,Y,Z,W) will
2864 correspond to consecutive words in the same order and format
2865 they're found in memory. No element-to-address conversion will be
2866 performed either: the value of the provided X coordinate will be
2867 interpreted in byte units instead of texel units. The result of
2868 accessing a misaligned address is undefined.
2870 Usage of the STORE opcode is only allowed if the WR (writable) flag
2875 ^^^^^^^^^^^^^^^^^^^^^^^^
2877 Properties are general directives that apply to the whole TGSI program.
2882 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
2883 The default value is UPPER_LEFT.
2885 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
2886 increase downward and rightward.
2887 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
2888 increase upward and rightward.
2890 OpenGL defaults to LOWER_LEFT, and is configurable with the
2891 GL_ARB_fragment_coord_conventions extension.
2893 DirectX 9/10 use UPPER_LEFT.
2895 FS_COORD_PIXEL_CENTER
2896 """""""""""""""""""""
2898 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
2899 The default value is HALF_INTEGER.
2901 If HALF_INTEGER, the fractionary part of the position will be 0.5
2902 If INTEGER, the fractionary part of the position will be 0.0
2904 Note that this does not affect the set of fragments generated by
2905 rasterization, which is instead controlled by half_pixel_center in the
2908 OpenGL defaults to HALF_INTEGER, and is configurable with the
2909 GL_ARB_fragment_coord_conventions extension.
2911 DirectX 9 uses INTEGER.
2912 DirectX 10 uses HALF_INTEGER.
2914 FS_COLOR0_WRITES_ALL_CBUFS
2915 """"""""""""""""""""""""""
2916 Specifies that writes to the fragment shader color 0 are replicated to all
2917 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
2918 fragData is directed to a single color buffer, but fragColor is broadcast.
2921 """"""""""""""""""""""""""
2922 If this property is set on the program bound to the shader stage before the
2923 fragment shader, user clip planes should have no effect (be disabled) even if
2924 that shader does not write to any clip distance outputs and the rasterizer's
2925 clip_plane_enable is non-zero.
2926 This property is only supported by drivers that also support shader clip
2928 This is useful for APIs that don't have UCPs and where clip distances written
2929 by a shader cannot be disabled.
2934 Specifies the number of times a geometry shader should be executed for each
2935 input primitive. Each invocation will have a different
2936 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
2939 VS_WINDOW_SPACE_POSITION
2940 """"""""""""""""""""""""""
2941 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
2942 is assumed to contain window space coordinates.
2943 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
2944 directly taken from the 4-th component of the shader output.
2945 Naturally, clipping is not performed on window coordinates either.
2946 The effect of this property is undefined if a geometry or tessellation shader
2949 Texture Sampling and Texture Formats
2950 ------------------------------------
2952 This table shows how texture image components are returned as (x,y,z,w) tuples
2953 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
2954 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
2957 +--------------------+--------------+--------------------+--------------+
2958 | Texture Components | Gallium | OpenGL | Direct3D 9 |
2959 +====================+==============+====================+==============+
2960 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
2961 +--------------------+--------------+--------------------+--------------+
2962 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
2963 +--------------------+--------------+--------------------+--------------+
2964 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
2965 +--------------------+--------------+--------------------+--------------+
2966 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
2967 +--------------------+--------------+--------------------+--------------+
2968 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
2969 +--------------------+--------------+--------------------+--------------+
2970 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
2971 +--------------------+--------------+--------------------+--------------+
2972 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
2973 +--------------------+--------------+--------------------+--------------+
2974 | I | (i, i, i, i) | (i, i, i, i) | N/A |
2975 +--------------------+--------------+--------------------+--------------+
2976 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
2977 | | | [#envmap-bumpmap]_ | |
2978 +--------------------+--------------+--------------------+--------------+
2979 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
2980 | | | [#depth-tex-mode]_ | |
2981 +--------------------+--------------+--------------------+--------------+
2982 | S | (s, s, s, s) | unknown | unknown |
2983 +--------------------+--------------+--------------------+--------------+
2985 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
2986 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
2987 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.