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 = (int) \lfloor src.x\rfloor
53 dst.y = (int) \lfloor src.y\rfloor
55 dst.z = (int) \lfloor src.z\rfloor
57 dst.w = (int) \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:: FMA - Fused Multiply-Add
277 Perform a * b + c with no intermediate rounding step.
281 dst.x = src0.x \times src1.x + src2.x
283 dst.y = src0.y \times src1.y + src2.y
285 dst.z = src0.z \times src1.z + src2.z
287 dst.w = src0.w \times src1.w + src2.w
290 .. opcode:: DP2A - 2-component Dot Product And Add
294 dst.x = src0.x \times src1.x + src0.y \times src1.y + src2.x
296 dst.y = src0.x \times src1.x + src0.y \times src1.y + src2.x
298 dst.z = src0.x \times src1.x + src0.y \times src1.y + src2.x
300 dst.w = src0.x \times src1.x + src0.y \times src1.y + src2.x
303 .. opcode:: FRC - Fraction
307 dst.x = src.x - \lfloor src.x\rfloor
309 dst.y = src.y - \lfloor src.y\rfloor
311 dst.z = src.z - \lfloor src.z\rfloor
313 dst.w = src.w - \lfloor src.w\rfloor
316 .. opcode:: CLAMP - Clamp
320 dst.x = clamp(src0.x, src1.x, src2.x)
322 dst.y = clamp(src0.y, src1.y, src2.y)
324 dst.z = clamp(src0.z, src1.z, src2.z)
326 dst.w = clamp(src0.w, src1.w, src2.w)
329 .. opcode:: FLR - Floor
333 dst.x = \lfloor src.x\rfloor
335 dst.y = \lfloor src.y\rfloor
337 dst.z = \lfloor src.z\rfloor
339 dst.w = \lfloor src.w\rfloor
342 .. opcode:: ROUND - Round
355 .. opcode:: EX2 - Exponential Base 2
357 This instruction replicates its result.
364 .. opcode:: LG2 - Logarithm Base 2
366 This instruction replicates its result.
373 .. opcode:: POW - Power
375 This instruction replicates its result.
379 dst = src0.x^{src1.x}
381 .. opcode:: XPD - Cross Product
385 dst.x = src0.y \times src1.z - src1.y \times src0.z
387 dst.y = src0.z \times src1.x - src1.z \times src0.x
389 dst.z = src0.x \times src1.y - src1.x \times src0.y
394 .. opcode:: ABS - Absolute
407 .. opcode:: DPH - Homogeneous Dot Product
409 This instruction replicates its result.
413 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
416 .. opcode:: COS - Cosine
418 This instruction replicates its result.
425 .. opcode:: DDX, DDX_FINE - Derivative Relative To X
427 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
428 advertised. When it is, the fine version guarantees one derivative per row
429 while DDX is allowed to be the same for the entire 2x2 quad.
433 dst.x = partialx(src.x)
435 dst.y = partialx(src.y)
437 dst.z = partialx(src.z)
439 dst.w = partialx(src.w)
442 .. opcode:: DDY, DDY_FINE - Derivative Relative To Y
444 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
445 advertised. When it is, the fine version guarantees one derivative per column
446 while DDY is allowed to be the same for the entire 2x2 quad.
450 dst.x = partialy(src.x)
452 dst.y = partialy(src.y)
454 dst.z = partialy(src.z)
456 dst.w = partialy(src.w)
459 .. opcode:: PK2H - Pack Two 16-bit Floats
464 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
469 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
474 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
479 .. opcode:: SEQ - Set On Equal
483 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
485 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
487 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
489 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
492 .. opcode:: SGT - Set On Greater Than
496 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
498 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
500 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
502 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
505 .. opcode:: SIN - Sine
507 This instruction replicates its result.
514 .. opcode:: SLE - Set On Less Equal Than
518 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
520 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
522 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
524 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
527 .. opcode:: SNE - Set On Not Equal
531 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
533 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
535 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
537 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
540 .. opcode:: TEX - Texture Lookup
542 for array textures src0.y contains the slice for 1D,
543 and src0.z contain the slice for 2D.
545 for shadow textures with no arrays (and not cube map),
546 src0.z contains the reference value.
548 for shadow textures with arrays, src0.z contains
549 the reference value for 1D arrays, and src0.w contains
550 the reference value for 2D arrays and cube maps.
552 for cube map array shadow textures, the reference value
553 cannot be passed in src0.w, and TEX2 must be used instead.
559 shadow_ref = src0.z or src0.w (optional)
563 dst = texture\_sample(unit, coord, shadow_ref)
566 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
568 this is the same as TEX, but uses another reg to encode the
579 dst = texture\_sample(unit, coord, shadow_ref)
584 .. opcode:: TXD - Texture Lookup with Derivatives
596 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
599 .. opcode:: TXP - Projective Texture Lookup
603 coord.x = src0.x / src0.w
605 coord.y = src0.y / src0.w
607 coord.z = src0.z / src0.w
613 dst = texture\_sample(unit, coord)
616 .. opcode:: UP2H - Unpack Two 16-Bit Floats
622 Considered for removal.
624 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
630 Considered for removal.
632 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
638 Considered for removal.
640 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
646 Considered for removal.
649 .. opcode:: ARR - Address Register Load With Round
653 dst.x = (int) round(src.x)
655 dst.y = (int) round(src.y)
657 dst.z = (int) round(src.z)
659 dst.w = (int) round(src.w)
662 .. opcode:: SSG - Set Sign
666 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
668 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
670 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
672 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
675 .. opcode:: CMP - Compare
679 dst.x = (src0.x < 0) ? src1.x : src2.x
681 dst.y = (src0.y < 0) ? src1.y : src2.y
683 dst.z = (src0.z < 0) ? src1.z : src2.z
685 dst.w = (src0.w < 0) ? src1.w : src2.w
688 .. opcode:: KILL_IF - Conditional Discard
690 Conditional discard. Allowed in fragment shaders only.
694 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
699 .. opcode:: KILL - Discard
701 Unconditional discard. Allowed in fragment shaders only.
704 .. opcode:: SCS - Sine Cosine
717 .. opcode:: TXB - Texture Lookup With Bias
719 for cube map array textures and shadow cube maps, the bias value
720 cannot be passed in src0.w, and TXB2 must be used instead.
722 if the target is a shadow texture, the reference value is always
723 in src.z (this prevents shadow 3d and shadow 2d arrays from
724 using this instruction, but this is not needed).
740 dst = texture\_sample(unit, coord, bias)
743 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
745 this is the same as TXB, but uses another reg to encode the
746 lod bias value for cube map arrays and shadow cube maps.
747 Presumably shadow 2d arrays and shadow 3d targets could use
748 this encoding too, but this is not legal.
750 shadow cube map arrays are neither possible nor required.
760 dst = texture\_sample(unit, coord, bias)
763 .. opcode:: DIV - Divide
767 dst.x = \frac{src0.x}{src1.x}
769 dst.y = \frac{src0.y}{src1.y}
771 dst.z = \frac{src0.z}{src1.z}
773 dst.w = \frac{src0.w}{src1.w}
776 .. opcode:: DP2 - 2-component Dot Product
778 This instruction replicates its result.
782 dst = src0.x \times src1.x + src0.y \times src1.y
785 .. opcode:: TXL - Texture Lookup With explicit LOD
787 for cube map array textures, the explicit lod value
788 cannot be passed in src0.w, and TXL2 must be used instead.
790 if the target is a shadow texture, the reference value is always
791 in src.z (this prevents shadow 3d / 2d array / cube targets from
792 using this instruction, but this is not needed).
808 dst = texture\_sample(unit, coord, lod)
811 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
813 this is the same as TXL, but uses another reg to encode the
815 Presumably shadow 3d / 2d array / cube targets could use
816 this encoding too, but this is not legal.
818 shadow cube map arrays are neither possible nor required.
828 dst = texture\_sample(unit, coord, lod)
831 .. opcode:: PUSHA - Push Address Register On Stack
840 Considered for cleanup.
844 Considered for removal.
846 .. opcode:: POPA - Pop Address Register From Stack
855 Considered for cleanup.
859 Considered for removal.
862 .. opcode:: CALLNZ - Subroutine Call If Not Zero
868 Considered for cleanup.
872 Considered for removal.
876 ^^^^^^^^^^^^^^^^^^^^^^^^
878 These opcodes are primarily provided for special-use computational shaders.
879 Support for these opcodes indicated by a special pipe capability bit (TBD).
881 XXX doesn't look like most of the opcodes really belong here.
883 .. opcode:: CEIL - Ceiling
887 dst.x = \lceil src.x\rceil
889 dst.y = \lceil src.y\rceil
891 dst.z = \lceil src.z\rceil
893 dst.w = \lceil src.w\rceil
896 .. opcode:: TRUNC - Truncate
909 .. opcode:: MOD - Modulus
913 dst.x = src0.x \bmod src1.x
915 dst.y = src0.y \bmod src1.y
917 dst.z = src0.z \bmod src1.z
919 dst.w = src0.w \bmod src1.w
922 .. opcode:: UARL - Integer Address Register Load
924 Moves the contents of the source register, assumed to be an integer, into the
925 destination register, which is assumed to be an address (ADDR) register.
928 .. opcode:: SAD - Sum Of Absolute Differences
932 dst.x = |src0.x - src1.x| + src2.x
934 dst.y = |src0.y - src1.y| + src2.y
936 dst.z = |src0.z - src1.z| + src2.z
938 dst.w = |src0.w - src1.w| + src2.w
941 .. opcode:: TXF - Texel Fetch
943 As per NV_gpu_shader4, extract a single texel from a specified texture
944 image. The source sampler may not be a CUBE or SHADOW. src 0 is a
945 four-component signed integer vector used to identify the single texel
946 accessed. 3 components + level. Just like texture instructions, an optional
947 offset vector is provided, which is subject to various driver restrictions
948 (regarding range, source of offsets).
949 TXF(uint_vec coord, int_vec offset).
952 .. opcode:: TXQ - Texture Size Query
954 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
955 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
956 depth), 1D array (width, layers), 2D array (width, height, layers).
957 Also return the number of accessible levels (last_level - first_level + 1)
960 For components which don't return a resource dimension, their value
967 dst.x = texture\_width(unit, lod)
969 dst.y = texture\_height(unit, lod)
971 dst.z = texture\_depth(unit, lod)
973 dst.w = texture\_levels(unit)
976 .. opcode:: TXQS - Texture Samples Query
978 This retrieves the number of samples in the texture, and stores it
979 into the x component. The other components are undefined.
983 dst.x = texture\_samples(unit)
986 .. opcode:: TG4 - Texture Gather
988 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
989 filtering operation and packs them into a single register. Only works with
990 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
991 addressing modes of the sampler and the top level of any mip pyramid are
992 used. Set W to zero. It behaves like the TEX instruction, but a filtered
993 sample is not generated. The four samples that contribute to filtering are
994 placed into xyzw in clockwise order, starting with the (u,v) texture
995 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
996 where the magnitude of the deltas are half a texel.
998 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
999 depth compares, single component selection, and a non-constant offset. It
1000 doesn't allow support for the GL independent offset to get i0,j0. This would
1001 require another CAP is hw can do it natively. For now we lower that before
1010 dst = texture\_gather4 (unit, coord, component)
1012 (with SM5 - cube array shadow)
1020 dst = texture\_gather (uint, coord, compare)
1022 .. opcode:: LODQ - level of detail query
1024 Compute the LOD information that the texture pipe would use to access the
1025 texture. The Y component contains the computed LOD lambda_prime. The X
1026 component contains the LOD that will be accessed, based on min/max lod's
1033 dst.xy = lodq(uint, coord);
1036 ^^^^^^^^^^^^^^^^^^^^^^^^
1037 These opcodes are used for integer operations.
1038 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1041 .. opcode:: I2F - Signed Integer To Float
1043 Rounding is unspecified (round to nearest even suggested).
1047 dst.x = (float) src.x
1049 dst.y = (float) src.y
1051 dst.z = (float) src.z
1053 dst.w = (float) src.w
1056 .. opcode:: U2F - Unsigned Integer To Float
1058 Rounding is unspecified (round to nearest even suggested).
1062 dst.x = (float) src.x
1064 dst.y = (float) src.y
1066 dst.z = (float) src.z
1068 dst.w = (float) src.w
1071 .. opcode:: F2I - Float to Signed Integer
1073 Rounding is towards zero (truncate).
1074 Values outside signed range (including NaNs) produce undefined results.
1087 .. opcode:: F2U - Float to Unsigned Integer
1089 Rounding is towards zero (truncate).
1090 Values outside unsigned range (including NaNs) produce undefined results.
1094 dst.x = (unsigned) src.x
1096 dst.y = (unsigned) src.y
1098 dst.z = (unsigned) src.z
1100 dst.w = (unsigned) src.w
1103 .. opcode:: UADD - Integer Add
1105 This instruction works the same for signed and unsigned integers.
1106 The low 32bit of the result is returned.
1110 dst.x = src0.x + src1.x
1112 dst.y = src0.y + src1.y
1114 dst.z = src0.z + src1.z
1116 dst.w = src0.w + src1.w
1119 .. opcode:: UMAD - Integer Multiply And Add
1121 This instruction works the same for signed and unsigned integers.
1122 The multiplication returns the low 32bit (as does the result itself).
1126 dst.x = src0.x \times src1.x + src2.x
1128 dst.y = src0.y \times src1.y + src2.y
1130 dst.z = src0.z \times src1.z + src2.z
1132 dst.w = src0.w \times src1.w + src2.w
1135 .. opcode:: UMUL - Integer Multiply
1137 This instruction works the same for signed and unsigned integers.
1138 The low 32bit of the result is returned.
1142 dst.x = src0.x \times src1.x
1144 dst.y = src0.y \times src1.y
1146 dst.z = src0.z \times src1.z
1148 dst.w = src0.w \times src1.w
1151 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1153 The high 32bits of the multiplication of 2 signed integers are returned.
1157 dst.x = (src0.x \times src1.x) >> 32
1159 dst.y = (src0.y \times src1.y) >> 32
1161 dst.z = (src0.z \times src1.z) >> 32
1163 dst.w = (src0.w \times src1.w) >> 32
1166 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1168 The high 32bits of the multiplication of 2 unsigned integers are returned.
1172 dst.x = (src0.x \times src1.x) >> 32
1174 dst.y = (src0.y \times src1.y) >> 32
1176 dst.z = (src0.z \times src1.z) >> 32
1178 dst.w = (src0.w \times src1.w) >> 32
1181 .. opcode:: IDIV - Signed Integer Division
1183 TBD: behavior for division by zero.
1187 dst.x = src0.x \ src1.x
1189 dst.y = src0.y \ src1.y
1191 dst.z = src0.z \ src1.z
1193 dst.w = src0.w \ src1.w
1196 .. opcode:: UDIV - Unsigned Integer Division
1198 For division by zero, 0xffffffff is returned.
1202 dst.x = src0.x \ src1.x
1204 dst.y = src0.y \ src1.y
1206 dst.z = src0.z \ src1.z
1208 dst.w = src0.w \ src1.w
1211 .. opcode:: UMOD - Unsigned Integer Remainder
1213 If second arg is zero, 0xffffffff is returned.
1217 dst.x = src0.x \ src1.x
1219 dst.y = src0.y \ src1.y
1221 dst.z = src0.z \ src1.z
1223 dst.w = src0.w \ src1.w
1226 .. opcode:: NOT - Bitwise Not
1239 .. opcode:: AND - Bitwise And
1243 dst.x = src0.x \& src1.x
1245 dst.y = src0.y \& src1.y
1247 dst.z = src0.z \& src1.z
1249 dst.w = src0.w \& src1.w
1252 .. opcode:: OR - Bitwise Or
1256 dst.x = src0.x | src1.x
1258 dst.y = src0.y | src1.y
1260 dst.z = src0.z | src1.z
1262 dst.w = src0.w | src1.w
1265 .. opcode:: XOR - Bitwise Xor
1269 dst.x = src0.x \oplus src1.x
1271 dst.y = src0.y \oplus src1.y
1273 dst.z = src0.z \oplus src1.z
1275 dst.w = src0.w \oplus src1.w
1278 .. opcode:: IMAX - Maximum of Signed Integers
1282 dst.x = max(src0.x, src1.x)
1284 dst.y = max(src0.y, src1.y)
1286 dst.z = max(src0.z, src1.z)
1288 dst.w = max(src0.w, src1.w)
1291 .. opcode:: UMAX - Maximum of Unsigned Integers
1295 dst.x = max(src0.x, src1.x)
1297 dst.y = max(src0.y, src1.y)
1299 dst.z = max(src0.z, src1.z)
1301 dst.w = max(src0.w, src1.w)
1304 .. opcode:: IMIN - Minimum of Signed Integers
1308 dst.x = min(src0.x, src1.x)
1310 dst.y = min(src0.y, src1.y)
1312 dst.z = min(src0.z, src1.z)
1314 dst.w = min(src0.w, src1.w)
1317 .. opcode:: UMIN - Minimum of Unsigned Integers
1321 dst.x = min(src0.x, src1.x)
1323 dst.y = min(src0.y, src1.y)
1325 dst.z = min(src0.z, src1.z)
1327 dst.w = min(src0.w, src1.w)
1330 .. opcode:: SHL - Shift Left
1332 The shift count is masked with 0x1f before the shift is applied.
1336 dst.x = src0.x << (0x1f \& src1.x)
1338 dst.y = src0.y << (0x1f \& src1.y)
1340 dst.z = src0.z << (0x1f \& src1.z)
1342 dst.w = src0.w << (0x1f \& src1.w)
1345 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1347 The shift count is masked with 0x1f before the shift is applied.
1351 dst.x = src0.x >> (0x1f \& src1.x)
1353 dst.y = src0.y >> (0x1f \& src1.y)
1355 dst.z = src0.z >> (0x1f \& src1.z)
1357 dst.w = src0.w >> (0x1f \& src1.w)
1360 .. opcode:: USHR - Logical Shift Right
1362 The shift count is masked with 0x1f before the shift is applied.
1366 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1368 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1370 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1372 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1375 .. opcode:: UCMP - Integer Conditional Move
1379 dst.x = src0.x ? src1.x : src2.x
1381 dst.y = src0.y ? src1.y : src2.y
1383 dst.z = src0.z ? src1.z : src2.z
1385 dst.w = src0.w ? src1.w : src2.w
1389 .. opcode:: ISSG - Integer Set Sign
1393 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1395 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1397 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1399 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1403 .. opcode:: FSLT - Float Set On Less Than (ordered)
1405 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1409 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1411 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1413 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1415 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1418 .. opcode:: ISLT - Signed Integer Set On Less Than
1422 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1424 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1426 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1428 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1431 .. opcode:: USLT - Unsigned Integer Set On Less Than
1435 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1437 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1439 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1441 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1444 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1446 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1450 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1452 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1454 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1456 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1459 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1463 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1465 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1467 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1469 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1472 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1476 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1478 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1480 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1482 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1485 .. opcode:: FSEQ - Float Set On Equal (ordered)
1487 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1491 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1493 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1495 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1497 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1500 .. opcode:: USEQ - Integer Set On Equal
1504 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1506 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1508 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1510 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1513 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1515 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1519 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1521 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1523 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1525 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1528 .. opcode:: USNE - Integer Set On Not Equal
1532 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1534 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1536 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1538 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1541 .. opcode:: INEG - Integer Negate
1556 .. opcode:: IABS - Integer Absolute Value
1570 These opcodes are used for bit-level manipulation of integers.
1572 .. opcode:: IBFE - Signed Bitfield Extract
1574 See SM5 instruction of the same name. Extracts a set of bits from the input,
1575 and sign-extends them if the high bit of the extracted window is set.
1579 def ibfe(value, offset, bits):
1580 offset = offset & 0x1f
1582 if bits == 0: return 0
1583 # Note: >> sign-extends
1584 if width + offset < 32:
1585 return (value << (32 - offset - bits)) >> (32 - bits)
1587 return value >> offset
1589 .. opcode:: UBFE - Unsigned Bitfield Extract
1591 See SM5 instruction of the same name. Extracts a set of bits from the input,
1592 without any sign-extension.
1596 def ubfe(value, offset, bits):
1597 offset = offset & 0x1f
1599 if bits == 0: return 0
1600 # Note: >> does not sign-extend
1601 if width + offset < 32:
1602 return (value << (32 - offset - bits)) >> (32 - bits)
1604 return value >> offset
1606 .. opcode:: BFI - Bitfield Insert
1608 See SM5 instruction of the same name. Replaces a bit region of 'base' with
1609 the low bits of 'insert'.
1613 def bfi(base, insert, offset, bits):
1614 offset = offset & 0x1f
1616 mask = ((1 << bits) - 1) << offset
1617 return ((insert << offset) & mask) | (base & ~mask)
1619 .. opcode:: BREV - Bitfield Reverse
1621 See SM5 instruction BFREV. Reverses the bits of the argument.
1623 .. opcode:: POPC - Population Count
1625 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1627 .. opcode:: LSB - Index of lowest set bit
1629 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1630 bit of the argument. Returns -1 if none are set.
1632 .. opcode:: IMSB - Index of highest non-sign bit
1634 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1635 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1636 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1637 (i.e. for inputs 0 and -1).
1639 .. opcode:: UMSB - Index of highest set bit
1641 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1642 set bit of the argument. Returns -1 if none are set.
1645 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1647 These opcodes are only supported in geometry shaders; they have no meaning
1648 in any other type of shader.
1650 .. opcode:: EMIT - Emit
1652 Generate a new vertex for the current primitive into the specified vertex
1653 stream using the values in the output registers.
1656 .. opcode:: ENDPRIM - End Primitive
1658 Complete the current primitive in the specified vertex stream (consisting of
1659 the emitted vertices), and start a new one.
1665 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1666 opcodes is determined by a special capability bit, ``GLSL``.
1667 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1669 .. opcode:: CAL - Subroutine Call
1675 .. opcode:: RET - Subroutine Call Return
1680 .. opcode:: CONT - Continue
1682 Unconditionally moves the point of execution to the instruction after the
1683 last bgnloop. The instruction must appear within a bgnloop/endloop.
1687 Support for CONT is determined by a special capability bit,
1688 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1691 .. opcode:: BGNLOOP - Begin a Loop
1693 Start a loop. Must have a matching endloop.
1696 .. opcode:: BGNSUB - Begin Subroutine
1698 Starts definition of a subroutine. Must have a matching endsub.
1701 .. opcode:: ENDLOOP - End a Loop
1703 End a loop started with bgnloop.
1706 .. opcode:: ENDSUB - End Subroutine
1708 Ends definition of a subroutine.
1711 .. opcode:: NOP - No Operation
1716 .. opcode:: BRK - Break
1718 Unconditionally moves the point of execution to the instruction after the
1719 next endloop or endswitch. The instruction must appear within a loop/endloop
1720 or switch/endswitch.
1723 .. opcode:: BREAKC - Break Conditional
1725 Conditionally moves the point of execution to the instruction after the
1726 next endloop or endswitch. The instruction must appear within a loop/endloop
1727 or switch/endswitch.
1728 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1729 as an integer register.
1733 Considered for removal as it's quite inconsistent wrt other opcodes
1734 (could emulate with UIF/BRK/ENDIF).
1737 .. opcode:: IF - Float If
1739 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1743 where src0.x is interpreted as a floating point register.
1746 .. opcode:: UIF - Bitwise If
1748 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1752 where src0.x is interpreted as an integer register.
1755 .. opcode:: ELSE - Else
1757 Starts an else block, after an IF or UIF statement.
1760 .. opcode:: ENDIF - End If
1762 Ends an IF or UIF block.
1765 .. opcode:: SWITCH - Switch
1767 Starts a C-style switch expression. The switch consists of one or multiple
1768 CASE statements, and at most one DEFAULT statement. Execution of a statement
1769 ends when a BRK is hit, but just like in C falling through to other cases
1770 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1771 just as last statement, and fallthrough is allowed into/from it.
1772 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1778 (some instructions here)
1781 (some instructions here)
1784 (some instructions here)
1789 .. opcode:: CASE - Switch case
1791 This represents a switch case label. The src arg must be an integer immediate.
1794 .. opcode:: DEFAULT - Switch default
1796 This represents the default case in the switch, which is taken if no other
1800 .. opcode:: ENDSWITCH - End of switch
1802 Ends a switch expression.
1808 The interpolation instructions allow an input to be interpolated in a
1809 different way than its declaration. This corresponds to the GLSL 4.00
1810 interpolateAt* functions. The first argument of each of these must come from
1811 ``TGSI_FILE_INPUT``.
1813 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1815 Interpolates the varying specified by src0 at the centroid
1817 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1819 Interpolates the varying specified by src0 at the sample id specified by
1820 src1.x (interpreted as an integer)
1822 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1824 Interpolates the varying specified by src0 at the offset src1.xy from the
1825 pixel center (interpreted as floats)
1833 The double-precision opcodes reinterpret four-component vectors into
1834 two-component vectors with doubled precision in each component.
1836 .. opcode:: DABS - Absolute
1841 .. opcode:: DADD - Add
1845 dst.xy = src0.xy + src1.xy
1847 dst.zw = src0.zw + src1.zw
1849 .. opcode:: DSEQ - Set on Equal
1853 dst.x = src0.xy == src1.xy ? \sim 0 : 0
1855 dst.z = src0.zw == src1.zw ? \sim 0 : 0
1857 .. opcode:: DSNE - Set on Equal
1861 dst.x = src0.xy != src1.xy ? \sim 0 : 0
1863 dst.z = src0.zw != src1.zw ? \sim 0 : 0
1865 .. opcode:: DSLT - Set on Less than
1869 dst.x = src0.xy < src1.xy ? \sim 0 : 0
1871 dst.z = src0.zw < src1.zw ? \sim 0 : 0
1873 .. opcode:: DSGE - Set on Greater equal
1877 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
1879 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
1881 .. opcode:: DFRAC - Fraction
1885 dst.xy = src.xy - \lfloor src.xy\rfloor
1887 dst.zw = src.zw - \lfloor src.zw\rfloor
1889 .. opcode:: DTRUNC - Truncate
1893 dst.xy = trunc(src.xy)
1895 dst.zw = trunc(src.zw)
1897 .. opcode:: DCEIL - Ceiling
1901 dst.xy = \lceil src.xy\rceil
1903 dst.zw = \lceil src.zw\rceil
1905 .. opcode:: DFLR - Floor
1909 dst.xy = \lfloor src.xy\rfloor
1911 dst.zw = \lfloor src.zw\rfloor
1913 .. opcode:: DROUND - Fraction
1917 dst.xy = round(src.xy)
1919 dst.zw = round(src.zw)
1921 .. opcode:: DSSG - Set Sign
1925 dst.xy = (src.xy > 0) ? 1.0 : (src.xy < 0) ? -1.0 : 0.0
1927 dst.zw = (src.zw > 0) ? 1.0 : (src.zw < 0) ? -1.0 : 0.0
1929 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1931 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1932 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1933 :math:`dst1 \times 2^{dst0} = src` .
1937 dst0.xy = exp(src.xy)
1939 dst1.xy = frac(src.xy)
1941 dst0.zw = exp(src.zw)
1943 dst1.zw = frac(src.zw)
1945 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1947 This opcode is the inverse of :opcode:`DFRACEXP`. The second
1948 source is an integer.
1952 dst.xy = src0.xy \times 2^{src1.x}
1954 dst.zw = src0.zw \times 2^{src1.y}
1956 .. opcode:: DMIN - Minimum
1960 dst.xy = min(src0.xy, src1.xy)
1962 dst.zw = min(src0.zw, src1.zw)
1964 .. opcode:: DMAX - Maximum
1968 dst.xy = max(src0.xy, src1.xy)
1970 dst.zw = max(src0.zw, src1.zw)
1972 .. opcode:: DMUL - Multiply
1976 dst.xy = src0.xy \times src1.xy
1978 dst.zw = src0.zw \times src1.zw
1981 .. opcode:: DMAD - Multiply And Add
1985 dst.xy = src0.xy \times src1.xy + src2.xy
1987 dst.zw = src0.zw \times src1.zw + src2.zw
1990 .. opcode:: DFMA - Fused Multiply-Add
1992 Perform a * b + c with no intermediate rounding step.
1996 dst.xy = src0.xy \times src1.xy + src2.xy
1998 dst.zw = src0.zw \times src1.zw + src2.zw
2001 .. opcode:: DRCP - Reciprocal
2005 dst.xy = \frac{1}{src.xy}
2007 dst.zw = \frac{1}{src.zw}
2009 .. opcode:: DSQRT - Square Root
2013 dst.xy = \sqrt{src.xy}
2015 dst.zw = \sqrt{src.zw}
2017 .. opcode:: DRSQ - Reciprocal Square Root
2021 dst.xy = \frac{1}{\sqrt{src.xy}}
2023 dst.zw = \frac{1}{\sqrt{src.zw}}
2025 .. opcode:: F2D - Float to Double
2029 dst.xy = double(src0.x)
2031 dst.zw = double(src0.y)
2033 .. opcode:: D2F - Double to Float
2037 dst.x = float(src0.xy)
2039 dst.y = float(src0.zw)
2041 .. opcode:: I2D - Int to Double
2045 dst.xy = double(src0.x)
2047 dst.zw = double(src0.y)
2049 .. opcode:: D2I - Double to Int
2053 dst.x = int(src0.xy)
2055 dst.y = int(src0.zw)
2057 .. opcode:: U2D - Unsigned Int to Double
2061 dst.xy = double(src0.x)
2063 dst.zw = double(src0.y)
2065 .. opcode:: D2U - Double to Unsigned Int
2069 dst.x = unsigned(src0.xy)
2071 dst.y = unsigned(src0.zw)
2073 .. _samplingopcodes:
2075 Resource Sampling Opcodes
2076 ^^^^^^^^^^^^^^^^^^^^^^^^^
2078 Those opcodes follow very closely semantics of the respective Direct3D
2079 instructions. If in doubt double check Direct3D documentation.
2080 Note that the swizzle on SVIEW (src1) determines texel swizzling
2085 Using provided address, sample data from the specified texture using the
2086 filtering mode identified by the gven sampler. The source data may come from
2087 any resource type other than buffers.
2089 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
2091 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
2093 .. opcode:: SAMPLE_I
2095 Simplified alternative to the SAMPLE instruction. Using the provided
2096 integer address, SAMPLE_I fetches data from the specified sampler view
2097 without any filtering. The source data may come from any resource type
2100 Syntax: ``SAMPLE_I dst, address, sampler_view``
2102 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
2104 The 'address' is specified as unsigned integers. If the 'address' is out of
2105 range [0...(# texels - 1)] the result of the fetch is always 0 in all
2106 components. As such the instruction doesn't honor address wrap modes, in
2107 cases where that behavior is desirable 'SAMPLE' instruction should be used.
2108 address.w always provides an unsigned integer mipmap level. If the value is
2109 out of the range then the instruction always returns 0 in all components.
2110 address.yz are ignored for buffers and 1d textures. address.z is ignored
2111 for 1d texture arrays and 2d textures.
2113 For 1D texture arrays address.y provides the array index (also as unsigned
2114 integer). If the value is out of the range of available array indices
2115 [0... (array size - 1)] then the opcode always returns 0 in all components.
2116 For 2D texture arrays address.z provides the array index, otherwise it
2117 exhibits the same behavior as in the case for 1D texture arrays. The exact
2118 semantics of the source address are presented in the table below:
2120 +---------------------------+----+-----+-----+---------+
2121 | resource type | X | Y | Z | W |
2122 +===========================+====+=====+=====+=========+
2123 | ``PIPE_BUFFER`` | x | | | ignored |
2124 +---------------------------+----+-----+-----+---------+
2125 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2126 +---------------------------+----+-----+-----+---------+
2127 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2128 +---------------------------+----+-----+-----+---------+
2129 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2130 +---------------------------+----+-----+-----+---------+
2131 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2132 +---------------------------+----+-----+-----+---------+
2133 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2134 +---------------------------+----+-----+-----+---------+
2135 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2136 +---------------------------+----+-----+-----+---------+
2137 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2138 +---------------------------+----+-----+-----+---------+
2140 Where 'mpl' is a mipmap level and 'idx' is the array index.
2142 .. opcode:: SAMPLE_I_MS
2144 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2146 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2148 .. opcode:: SAMPLE_B
2150 Just like the SAMPLE instruction with the exception that an additional bias
2151 is applied to the level of detail computed as part of the instruction
2154 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2156 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2158 .. opcode:: SAMPLE_C
2160 Similar to the SAMPLE instruction but it performs a comparison filter. The
2161 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2162 additional float32 operand, reference value, which must be a register with
2163 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2164 current samplers compare_func (in pipe_sampler_state) to compare reference
2165 value against the red component value for the surce resource at each texel
2166 that the currently configured texture filter covers based on the provided
2169 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2171 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2173 .. opcode:: SAMPLE_C_LZ
2175 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2178 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2180 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2183 .. opcode:: SAMPLE_D
2185 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2186 the source address in the x direction and the y direction are provided by
2189 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2191 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2193 .. opcode:: SAMPLE_L
2195 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2196 directly as a scalar value, representing no anisotropy.
2198 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2200 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2204 Gathers the four texels to be used in a bi-linear filtering operation and
2205 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2206 and cubemaps arrays. For 2D textures, only the addressing modes of the
2207 sampler and the top level of any mip pyramid are used. Set W to zero. It
2208 behaves like the SAMPLE instruction, but a filtered sample is not
2209 generated. The four samples that contribute to filtering are placed into
2210 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2211 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2212 magnitude of the deltas are half a texel.
2215 .. opcode:: SVIEWINFO
2217 Query the dimensions of a given sampler view. dst receives width, height,
2218 depth or array size and number of mipmap levels as int4. The dst can have a
2219 writemask which will specify what info is the caller interested in.
2221 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2223 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2225 src_mip_level is an unsigned integer scalar. If it's out of range then
2226 returns 0 for width, height and depth/array size but the total number of
2227 mipmap is still returned correctly for the given sampler view. The returned
2228 width, height and depth values are for the mipmap level selected by the
2229 src_mip_level and are in the number of texels. For 1d texture array width
2230 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2231 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2232 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2233 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2234 resinfo allowing swizzling dst values is ignored (due to the interaction
2235 with rcpfloat modifier which requires some swizzle handling in the state
2238 .. opcode:: SAMPLE_POS
2240 Query the position of a given sample. dst receives float4 (x, y, 0, 0)
2241 indicated where the sample is located. If the resource is not a multi-sample
2242 resource and not a render target, the result is 0.
2244 .. opcode:: SAMPLE_INFO
2246 dst receives number of samples in x. If the resource is not a multi-sample
2247 resource and not a render target, the result is 0.
2250 .. _resourceopcodes:
2252 Resource Access Opcodes
2253 ^^^^^^^^^^^^^^^^^^^^^^^
2255 .. opcode:: LOAD - Fetch data from a shader resource
2257 Syntax: ``LOAD dst, resource, address``
2259 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
2261 Using the provided integer address, LOAD fetches data
2262 from the specified buffer or texture without any
2265 The 'address' is specified as a vector of unsigned
2266 integers. If the 'address' is out of range the result
2269 Only the first mipmap level of a resource can be read
2270 from using this instruction.
2272 For 1D or 2D texture arrays, the array index is
2273 provided as an unsigned integer in address.y or
2274 address.z, respectively. address.yz are ignored for
2275 buffers and 1D textures. address.z is ignored for 1D
2276 texture arrays and 2D textures. address.w is always
2279 .. opcode:: STORE - Write data to a shader resource
2281 Syntax: ``STORE resource, address, src``
2283 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2285 Using the provided integer address, STORE writes data
2286 to the specified buffer or texture.
2288 The 'address' is specified as a vector of unsigned
2289 integers. If the 'address' is out of range the result
2292 Only the first mipmap level of a resource can be
2293 written to using this instruction.
2295 For 1D or 2D texture arrays, the array index is
2296 provided as an unsigned integer in address.y or
2297 address.z, respectively. address.yz are ignored for
2298 buffers and 1D textures. address.z is ignored for 1D
2299 texture arrays and 2D textures. address.w is always
2303 .. _threadsyncopcodes:
2305 Inter-thread synchronization opcodes
2306 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2308 These opcodes are intended for communication between threads running
2309 within the same compute grid. For now they're only valid in compute
2312 .. opcode:: MFENCE - Memory fence
2314 Syntax: ``MFENCE resource``
2316 Example: ``MFENCE RES[0]``
2318 This opcode forces strong ordering between any memory access
2319 operations that affect the specified resource. This means that
2320 previous loads and stores (and only those) will be performed and
2321 visible to other threads before the program execution continues.
2324 .. opcode:: LFENCE - Load memory fence
2326 Syntax: ``LFENCE resource``
2328 Example: ``LFENCE RES[0]``
2330 Similar to MFENCE, but it only affects the ordering of memory loads.
2333 .. opcode:: SFENCE - Store memory fence
2335 Syntax: ``SFENCE resource``
2337 Example: ``SFENCE RES[0]``
2339 Similar to MFENCE, but it only affects the ordering of memory stores.
2342 .. opcode:: BARRIER - Thread group barrier
2346 This opcode suspends the execution of the current thread until all
2347 the remaining threads in the working group reach the same point of
2348 the program. Results are unspecified if any of the remaining
2349 threads terminates or never reaches an executed BARRIER instruction.
2357 These opcodes provide atomic variants of some common arithmetic and
2358 logical operations. In this context atomicity means that another
2359 concurrent memory access operation that affects the same memory
2360 location is guaranteed to be performed strictly before or after the
2361 entire execution of the atomic operation.
2363 For the moment they're only valid in compute programs.
2365 .. opcode:: ATOMUADD - Atomic integer addition
2367 Syntax: ``ATOMUADD dst, resource, offset, src``
2369 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2371 The following operation is performed atomically on each component:
2375 dst_i = resource[offset]_i
2377 resource[offset]_i = dst_i + src_i
2380 .. opcode:: ATOMXCHG - Atomic exchange
2382 Syntax: ``ATOMXCHG dst, resource, offset, src``
2384 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2386 The following operation is performed atomically on each component:
2390 dst_i = resource[offset]_i
2392 resource[offset]_i = src_i
2395 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2397 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2399 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2401 The following operation is performed atomically on each component:
2405 dst_i = resource[offset]_i
2407 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2410 .. opcode:: ATOMAND - Atomic bitwise And
2412 Syntax: ``ATOMAND dst, resource, offset, src``
2414 Example: ``ATOMAND TEMP[0], RES[0], TEMP[1], TEMP[2]``
2416 The following operation is performed atomically on each component:
2420 dst_i = resource[offset]_i
2422 resource[offset]_i = dst_i \& src_i
2425 .. opcode:: ATOMOR - Atomic bitwise Or
2427 Syntax: ``ATOMOR dst, resource, offset, src``
2429 Example: ``ATOMOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2431 The following operation is performed atomically on each component:
2435 dst_i = resource[offset]_i
2437 resource[offset]_i = dst_i | src_i
2440 .. opcode:: ATOMXOR - Atomic bitwise Xor
2442 Syntax: ``ATOMXOR dst, resource, offset, src``
2444 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2446 The following operation is performed atomically on each component:
2450 dst_i = resource[offset]_i
2452 resource[offset]_i = dst_i \oplus src_i
2455 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2457 Syntax: ``ATOMUMIN dst, resource, offset, src``
2459 Example: ``ATOMUMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2461 The following operation is performed atomically on each component:
2465 dst_i = resource[offset]_i
2467 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2470 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2472 Syntax: ``ATOMUMAX dst, resource, offset, src``
2474 Example: ``ATOMUMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2476 The following operation is performed atomically on each component:
2480 dst_i = resource[offset]_i
2482 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2485 .. opcode:: ATOMIMIN - Atomic signed minimum
2487 Syntax: ``ATOMIMIN dst, resource, offset, src``
2489 Example: ``ATOMIMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2491 The following operation is performed atomically on each component:
2495 dst_i = resource[offset]_i
2497 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2500 .. opcode:: ATOMIMAX - Atomic signed maximum
2502 Syntax: ``ATOMIMAX dst, resource, offset, src``
2504 Example: ``ATOMIMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2506 The following operation is performed atomically on each component:
2510 dst_i = resource[offset]_i
2512 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2516 Explanation of symbols used
2517 ------------------------------
2524 :math:`|x|` Absolute value of `x`.
2526 :math:`\lceil x \rceil` Ceiling of `x`.
2528 clamp(x,y,z) Clamp x between y and z.
2529 (x < y) ? y : (x > z) ? z : x
2531 :math:`\lfloor x\rfloor` Floor of `x`.
2533 :math:`\log_2{x}` Logarithm of `x`, base 2.
2535 max(x,y) Maximum of x and y.
2538 min(x,y) Minimum of x and y.
2541 partialx(x) Derivative of x relative to fragment's X.
2543 partialy(x) Derivative of x relative to fragment's Y.
2545 pop() Pop from stack.
2547 :math:`x^y` `x` to the power `y`.
2549 push(x) Push x on stack.
2553 trunc(x) Truncate x, i.e. drop the fraction bits.
2560 discard Discard fragment.
2564 target Label of target instruction.
2575 Declares a register that is will be referenced as an operand in Instruction
2578 File field contains register file that is being declared and is one
2581 UsageMask field specifies which of the register components can be accessed
2582 and is one of TGSI_WRITEMASK.
2584 The Local flag specifies that a given value isn't intended for
2585 subroutine parameter passing and, as a result, the implementation
2586 isn't required to give any guarantees of it being preserved across
2587 subroutine boundaries. As it's merely a compiler hint, the
2588 implementation is free to ignore it.
2590 If Dimension flag is set to 1, a Declaration Dimension token follows.
2592 If Semantic flag is set to 1, a Declaration Semantic token follows.
2594 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2596 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2598 If Array flag is set to 1, a Declaration Array token follows.
2601 ^^^^^^^^^^^^^^^^^^^^^^^^
2603 Declarations can optional have an ArrayID attribute which can be referred by
2604 indirect addressing operands. An ArrayID of zero is reserved and treated as
2605 if no ArrayID is specified.
2607 If an indirect addressing operand refers to a specific declaration by using
2608 an ArrayID only the registers in this declaration are guaranteed to be
2609 accessed, accessing any register outside this declaration results in undefined
2610 behavior. Note that for compatibility the effective index is zero-based and
2611 not relative to the specified declaration
2613 If no ArrayID is specified with an indirect addressing operand the whole
2614 register file might be accessed by this operand. This is strongly discouraged
2615 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2616 This is only legal for TEMP and CONST register files.
2618 Declaration Semantic
2619 ^^^^^^^^^^^^^^^^^^^^^^^^
2621 Vertex and fragment shader input and output registers may be labeled
2622 with semantic information consisting of a name and index.
2624 Follows Declaration token if Semantic bit is set.
2626 Since its purpose is to link a shader with other stages of the pipeline,
2627 it is valid to follow only those Declaration tokens that declare a register
2628 either in INPUT or OUTPUT file.
2630 SemanticName field contains the semantic name of the register being declared.
2631 There is no default value.
2633 SemanticIndex is an optional subscript that can be used to distinguish
2634 different register declarations with the same semantic name. The default value
2637 The meanings of the individual semantic names are explained in the following
2640 TGSI_SEMANTIC_POSITION
2641 """"""""""""""""""""""
2643 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2644 output register which contains the homogeneous vertex position in the clip
2645 space coordinate system. After clipping, the X, Y and Z components of the
2646 vertex will be divided by the W value to get normalized device coordinates.
2648 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2649 fragment shader input contains the fragment's window position. The X
2650 component starts at zero and always increases from left to right.
2651 The Y component starts at zero and always increases but Y=0 may either
2652 indicate the top of the window or the bottom depending on the fragment
2653 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2654 The Z coordinate ranges from 0 to 1 to represent depth from the front
2655 to the back of the Z buffer. The W component contains the interpolated
2656 reciprocal of the vertex position W component (corresponding to gl_Fragcoord,
2657 but unlike d3d10 which interpolates the same 1/w but then gives back
2658 the reciprocal of the interpolated value).
2660 Fragment shaders may also declare an output register with
2661 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2662 the fragment shader to change the fragment's Z position.
2669 For vertex shader outputs or fragment shader inputs/outputs, this
2670 label indicates that the resister contains an R,G,B,A color.
2672 Several shader inputs/outputs may contain colors so the semantic index
2673 is used to distinguish them. For example, color[0] may be the diffuse
2674 color while color[1] may be the specular color.
2676 This label is needed so that the flat/smooth shading can be applied
2677 to the right interpolants during rasterization.
2681 TGSI_SEMANTIC_BCOLOR
2682 """"""""""""""""""""
2684 Back-facing colors are only used for back-facing polygons, and are only valid
2685 in vertex shader outputs. After rasterization, all polygons are front-facing
2686 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2687 so all BCOLORs effectively become regular COLORs in the fragment shader.
2693 Vertex shader inputs and outputs and fragment shader inputs may be
2694 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2695 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2696 to compute a fog blend factor which is used to blend the normal fragment color
2697 with a constant fog color. But fog coord really is just an ordinary vec4
2698 register like regular semantics.
2704 Vertex shader input and output registers may be labeled with
2705 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2706 in the form (S, 0, 0, 1). The point size controls the width or diameter
2707 of points for rasterization. This label cannot be used in fragment
2710 When using this semantic, be sure to set the appropriate state in the
2711 :ref:`rasterizer` first.
2714 TGSI_SEMANTIC_TEXCOORD
2715 """"""""""""""""""""""
2717 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2719 Vertex shader outputs and fragment shader inputs may be labeled with
2720 this semantic to make them replaceable by sprite coordinates via the
2721 sprite_coord_enable state in the :ref:`rasterizer`.
2722 The semantic index permitted with this semantic is limited to <= 7.
2724 If the driver does not support TEXCOORD, sprite coordinate replacement
2725 applies to inputs with the GENERIC semantic instead.
2727 The intended use case for this semantic is gl_TexCoord.
2730 TGSI_SEMANTIC_PCOORD
2731 """"""""""""""""""""
2733 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2735 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2736 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2737 the current primitive is a point and point sprites are enabled. Otherwise,
2738 the contents of the register are undefined.
2740 The intended use case for this semantic is gl_PointCoord.
2743 TGSI_SEMANTIC_GENERIC
2744 """""""""""""""""""""
2746 All vertex/fragment shader inputs/outputs not labeled with any other
2747 semantic label can be considered to be generic attributes. Typical
2748 uses of generic inputs/outputs are texcoords and user-defined values.
2751 TGSI_SEMANTIC_NORMAL
2752 """"""""""""""""""""
2754 Indicates that a vertex shader input is a normal vector. This is
2755 typically only used for legacy graphics APIs.
2761 This label applies to fragment shader inputs only and indicates that
2762 the register contains front/back-face information of the form (F, 0,
2763 0, 1). The first component will be positive when the fragment belongs
2764 to a front-facing polygon, and negative when the fragment belongs to a
2765 back-facing polygon.
2768 TGSI_SEMANTIC_EDGEFLAG
2769 """"""""""""""""""""""
2771 For vertex shaders, this sematic label indicates that an input or
2772 output is a boolean edge flag. The register layout is [F, x, x, x]
2773 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2774 simply copies the edge flag input to the edgeflag output.
2776 Edge flags are used to control which lines or points are actually
2777 drawn when the polygon mode converts triangles/quads/polygons into
2781 TGSI_SEMANTIC_STENCIL
2782 """""""""""""""""""""
2784 For fragment shaders, this semantic label indicates that an output
2785 is a writable stencil reference value. Only the Y component is writable.
2786 This allows the fragment shader to change the fragments stencilref value.
2789 TGSI_SEMANTIC_VIEWPORT_INDEX
2790 """"""""""""""""""""""""""""
2792 For geometry shaders, this semantic label indicates that an output
2793 contains the index of the viewport (and scissor) to use.
2794 This is an integer value, and only the X component is used.
2800 For geometry shaders, this semantic label indicates that an output
2801 contains the layer value to use for the color and depth/stencil surfaces.
2802 This is an integer value, and only the X component is used.
2803 (Also known as rendertarget array index.)
2806 TGSI_SEMANTIC_CULLDIST
2807 """"""""""""""""""""""
2809 Used as distance to plane for performing application-defined culling
2810 of individual primitives against a plane. When components of vertex
2811 elements are given this label, these values are assumed to be a
2812 float32 signed distance to a plane. Primitives will be completely
2813 discarded if the plane distance for all of the vertices in the
2814 primitive are < 0. If a vertex has a cull distance of NaN, that
2815 vertex counts as "out" (as if its < 0);
2816 The limits on both clip and cull distances are bound
2817 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2818 the maximum number of components that can be used to hold the
2819 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2820 which specifies the maximum number of registers which can be
2821 annotated with those semantics.
2824 TGSI_SEMANTIC_CLIPDIST
2825 """"""""""""""""""""""
2827 When components of vertex elements are identified this way, these
2828 values are each assumed to be a float32 signed distance to a plane.
2829 Primitive setup only invokes rasterization on pixels for which
2830 the interpolated plane distances are >= 0. Multiple clip planes
2831 can be implemented simultaneously, by annotating multiple
2832 components of one or more vertex elements with the above specified
2833 semantic. The limits on both clip and cull distances are bound
2834 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2835 the maximum number of components that can be used to hold the
2836 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2837 which specifies the maximum number of registers which can be
2838 annotated with those semantics.
2840 TGSI_SEMANTIC_SAMPLEID
2841 """"""""""""""""""""""
2843 For fragment shaders, this semantic label indicates that a system value
2844 contains the current sample id (i.e. gl_SampleID).
2845 This is an integer value, and only the X component is used.
2847 TGSI_SEMANTIC_SAMPLEPOS
2848 """""""""""""""""""""""
2850 For fragment shaders, this semantic label indicates that a system value
2851 contains the current sample's position (i.e. gl_SamplePosition). Only the X
2852 and Y values are used.
2854 TGSI_SEMANTIC_SAMPLEMASK
2855 """"""""""""""""""""""""
2857 For fragment shaders, this semantic label indicates that an output contains
2858 the sample mask used to disable further sample processing
2859 (i.e. gl_SampleMask). Only the X value is used, up to 32x MS.
2861 TGSI_SEMANTIC_INVOCATIONID
2862 """"""""""""""""""""""""""
2864 For geometry shaders, this semantic label indicates that a system value
2865 contains the current invocation id (i.e. gl_InvocationID).
2866 This is an integer value, and only the X component is used.
2868 TGSI_SEMANTIC_INSTANCEID
2869 """"""""""""""""""""""""
2871 For vertex shaders, this semantic label indicates that a system value contains
2872 the current instance id (i.e. gl_InstanceID). It does not include the base
2873 instance. This is an integer value, and only the X component is used.
2875 TGSI_SEMANTIC_VERTEXID
2876 """"""""""""""""""""""
2878 For vertex shaders, this semantic label indicates that a system value contains
2879 the current vertex id (i.e. gl_VertexID). It does (unlike in d3d10) include the
2880 base vertex. This is an integer value, and only the X component is used.
2882 TGSI_SEMANTIC_VERTEXID_NOBASE
2883 """""""""""""""""""""""""""""""
2885 For vertex shaders, this semantic label indicates that a system value contains
2886 the current vertex id without including the base vertex (this corresponds to
2887 d3d10 vertex id, so TGSI_SEMANTIC_VERTEXID_NOBASE + TGSI_SEMANTIC_BASEVERTEX
2888 == TGSI_SEMANTIC_VERTEXID). This is an integer value, and only the X component
2891 TGSI_SEMANTIC_BASEVERTEX
2892 """"""""""""""""""""""""
2894 For vertex shaders, this semantic label indicates that a system value contains
2895 the base vertex (i.e. gl_BaseVertex). Note that for non-indexed draw calls,
2896 this contains the first (or start) value instead.
2897 This is an integer value, and only the X component is used.
2899 TGSI_SEMANTIC_PRIMID
2900 """"""""""""""""""""
2902 For geometry and fragment shaders, this semantic label indicates the value
2903 contains the primitive id (i.e. gl_PrimitiveID). This is an integer value,
2904 and only the X component is used.
2905 FIXME: This right now can be either a ordinary input or a system value...
2911 For tessellation evaluation/control shaders, this semantic label indicates a
2912 generic per-patch attribute. Such semantics will not implicitly be per-vertex
2915 TGSI_SEMANTIC_TESSCOORD
2916 """""""""""""""""""""""
2918 For tessellation evaluation shaders, this semantic label indicates the
2919 coordinates of the vertex being processed. This is available in XYZ; W is
2922 TGSI_SEMANTIC_TESSOUTER
2923 """""""""""""""""""""""
2925 For tessellation evaluation/control shaders, this semantic label indicates the
2926 outer tessellation levels of the patch. Isoline tessellation will only have XY
2927 defined, triangle will have XYZ and quads will have XYZW defined. This
2928 corresponds to gl_TessLevelOuter.
2930 TGSI_SEMANTIC_TESSINNER
2931 """""""""""""""""""""""
2933 For tessellation evaluation/control shaders, this semantic label indicates the
2934 inner tessellation levels of the patch. The X value is only defined for
2935 triangle tessellation, while quads will have XY defined. This is entirely
2936 undefined for isoline tessellation.
2938 TGSI_SEMANTIC_VERTICESIN
2939 """"""""""""""""""""""""
2941 For tessellation evaluation/control shaders, this semantic label indicates the
2942 number of vertices provided in the input patch. Only the X value is defined.
2944 TGSI_SEMANTIC_HELPER_INVOCATION
2945 """""""""""""""""""""""""""""""
2947 For fragment shaders, this semantic indicates whether the current
2948 invocation is covered or not. Helper invocations are created in order
2949 to properly compute derivatives, however it may be desirable to skip
2950 some of the logic in those cases. See ``gl_HelperInvocation`` documentation.
2952 TGSI_SEMANTIC_BASEINSTANCE
2953 """"""""""""""""""""""""""
2955 For vertex shaders, the base instance argument supplied for this
2956 draw. This is an integer value, and only the X component is used.
2958 TGSI_SEMANTIC_DRAWID
2959 """"""""""""""""""""
2961 For vertex shaders, the zero-based index of the current draw in a
2962 ``glMultiDraw*`` invocation. This is an integer value, and only the X
2966 Declaration Interpolate
2967 ^^^^^^^^^^^^^^^^^^^^^^^
2969 This token is only valid for fragment shader INPUT declarations.
2971 The Interpolate field specifes the way input is being interpolated by
2972 the rasteriser and is one of TGSI_INTERPOLATE_*.
2974 The Location field specifies the location inside the pixel that the
2975 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
2976 when per-sample shading is enabled, the implementation may choose to
2977 interpolate at the sample irrespective of the Location field.
2979 The CylindricalWrap bitfield specifies which register components
2980 should be subject to cylindrical wrapping when interpolating by the
2981 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2982 should be interpolated according to cylindrical wrapping rules.
2985 Declaration Sampler View
2986 ^^^^^^^^^^^^^^^^^^^^^^^^
2988 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2990 DCL SVIEW[#], resource, type(s)
2992 Declares a shader input sampler view and assigns it to a SVIEW[#]
2995 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2997 type must be 1 or 4 entries (if specifying on a per-component
2998 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
3000 For TEX\* style texture sample opcodes (as opposed to SAMPLE\* opcodes
3001 which take an explicit SVIEW[#] source register), there may be optionally
3002 SVIEW[#] declarations. In this case, the SVIEW index is implied by the
3003 SAMP index, and there must be a corresponding SVIEW[#] declaration for
3004 each SAMP[#] declaration. Drivers are free to ignore this if they wish.
3005 But note in particular that some drivers need to know the sampler type
3006 (float/int/unsigned) in order to generate the correct code, so cases
3007 where integer textures are sampled, SVIEW[#] declarations should be
3010 NOTE: It is NOT legal to mix SAMPLE\* style opcodes and TEX\* opcodes
3013 Declaration Resource
3014 ^^^^^^^^^^^^^^^^^^^^
3016 Follows Declaration token if file is TGSI_FILE_RESOURCE.
3018 DCL RES[#], resource [, WR] [, RAW]
3020 Declares a shader input resource and assigns it to a RES[#]
3023 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
3026 If the RAW keyword is not specified, the texture data will be
3027 subject to conversion, swizzling and scaling as required to yield
3028 the specified data type from the physical data format of the bound
3031 If the RAW keyword is specified, no channel conversion will be
3032 performed: the values read for each of the channels (X,Y,Z,W) will
3033 correspond to consecutive words in the same order and format
3034 they're found in memory. No element-to-address conversion will be
3035 performed either: the value of the provided X coordinate will be
3036 interpreted in byte units instead of texel units. The result of
3037 accessing a misaligned address is undefined.
3039 Usage of the STORE opcode is only allowed if the WR (writable) flag
3044 ^^^^^^^^^^^^^^^^^^^^^^^^
3046 Properties are general directives that apply to the whole TGSI program.
3051 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
3052 The default value is UPPER_LEFT.
3054 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
3055 increase downward and rightward.
3056 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
3057 increase upward and rightward.
3059 OpenGL defaults to LOWER_LEFT, and is configurable with the
3060 GL_ARB_fragment_coord_conventions extension.
3062 DirectX 9/10 use UPPER_LEFT.
3064 FS_COORD_PIXEL_CENTER
3065 """""""""""""""""""""
3067 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
3068 The default value is HALF_INTEGER.
3070 If HALF_INTEGER, the fractionary part of the position will be 0.5
3071 If INTEGER, the fractionary part of the position will be 0.0
3073 Note that this does not affect the set of fragments generated by
3074 rasterization, which is instead controlled by half_pixel_center in the
3077 OpenGL defaults to HALF_INTEGER, and is configurable with the
3078 GL_ARB_fragment_coord_conventions extension.
3080 DirectX 9 uses INTEGER.
3081 DirectX 10 uses HALF_INTEGER.
3083 FS_COLOR0_WRITES_ALL_CBUFS
3084 """"""""""""""""""""""""""
3085 Specifies that writes to the fragment shader color 0 are replicated to all
3086 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
3087 fragData is directed to a single color buffer, but fragColor is broadcast.
3090 """"""""""""""""""""""""""
3091 If this property is set on the program bound to the shader stage before the
3092 fragment shader, user clip planes should have no effect (be disabled) even if
3093 that shader does not write to any clip distance outputs and the rasterizer's
3094 clip_plane_enable is non-zero.
3095 This property is only supported by drivers that also support shader clip
3097 This is useful for APIs that don't have UCPs and where clip distances written
3098 by a shader cannot be disabled.
3103 Specifies the number of times a geometry shader should be executed for each
3104 input primitive. Each invocation will have a different
3105 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
3108 VS_WINDOW_SPACE_POSITION
3109 """"""""""""""""""""""""""
3110 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
3111 is assumed to contain window space coordinates.
3112 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
3113 directly taken from the 4-th component of the shader output.
3114 Naturally, clipping is not performed on window coordinates either.
3115 The effect of this property is undefined if a geometry or tessellation shader
3121 The number of vertices written by the tessellation control shader. This
3122 effectively defines the patch input size of the tessellation evaluation shader
3128 This sets the tessellation primitive mode, one of ``PIPE_PRIM_TRIANGLES``,
3129 ``PIPE_PRIM_QUADS``, or ``PIPE_PRIM_LINES``. (Unlike in GL, there is no
3130 separate isolines settings, the regular lines is assumed to mean isolines.)
3135 This sets the spacing mode of the tessellation generator, one of
3136 ``PIPE_TESS_SPACING_*``.
3141 This sets the vertex order to be clockwise if the value is 1, or
3142 counter-clockwise if set to 0.
3147 If set to a non-zero value, this turns on point mode for the tessellator,
3148 which means that points will be generated instead of primitives.
3150 NUM_CLIPDIST_ENABLED
3153 How many clip distance scalar outputs are enabled.
3155 NUM_CULLDIST_ENABLED
3158 How many cull distance scalar outputs are enabled.
3161 Texture Sampling and Texture Formats
3162 ------------------------------------
3164 This table shows how texture image components are returned as (x,y,z,w) tuples
3165 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
3166 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
3169 +--------------------+--------------+--------------------+--------------+
3170 | Texture Components | Gallium | OpenGL | Direct3D 9 |
3171 +====================+==============+====================+==============+
3172 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
3173 +--------------------+--------------+--------------------+--------------+
3174 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
3175 +--------------------+--------------+--------------------+--------------+
3176 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
3177 +--------------------+--------------+--------------------+--------------+
3178 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
3179 +--------------------+--------------+--------------------+--------------+
3180 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
3181 +--------------------+--------------+--------------------+--------------+
3182 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
3183 +--------------------+--------------+--------------------+--------------+
3184 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
3185 +--------------------+--------------+--------------------+--------------+
3186 | I | (i, i, i, i) | (i, i, i, i) | N/A |
3187 +--------------------+--------------+--------------------+--------------+
3188 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
3189 | | | [#envmap-bumpmap]_ | |
3190 +--------------------+--------------+--------------------+--------------+
3191 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
3192 | | | [#depth-tex-mode]_ | |
3193 +--------------------+--------------+--------------------+--------------+
3194 | S | (s, s, s, s) | unknown | unknown |
3195 +--------------------+--------------+--------------------+--------------+
3197 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
3198 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
3199 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.