4 TGSI, Tungsten Graphics Shader Infrastructure, is an intermediate language
5 for describing shaders. Since Gallium is inherently shaderful, shaders are
6 an important part of the API. TGSI is the only intermediate representation
12 All TGSI instructions, known as *opcodes*, operate on arbitrary-precision
13 floating-point four-component vectors. An opcode may have up to one
14 destination register, known as *dst*, and between zero and three source
15 registers, called *src0* through *src2*, or simply *src* if there is only
18 Some instructions, like :opcode:`I2F`, permit re-interpretation of vector
19 components as integers. Other instructions permit using registers as
20 two-component vectors with double precision; see :ref:`doubleopcodes`.
22 When an instruction has a scalar result, the result is usually copied into
23 each of the components of *dst*. When this happens, the result is said to be
24 *replicated* to *dst*. :opcode:`RCP` is one such instruction.
29 TGSI supports modifiers on inputs (as well as saturate modifier on instructions).
31 For inputs which have a floating point type, both absolute value and negation
32 modifiers are supported (with absolute value being applied first).
33 TGSI_OPCODE_MOV is considered to have float input type for applying modifiers.
35 For inputs which have signed or unsigned type only the negate modifier is
42 ^^^^^^^^^^^^^^^^^^^^^^^^^
44 These opcodes are guaranteed to be available regardless of the driver being
47 .. opcode:: ARL - Address Register Load
51 dst.x = \lfloor src.x\rfloor
53 dst.y = \lfloor src.y\rfloor
55 dst.z = \lfloor src.z\rfloor
57 dst.w = \lfloor src.w\rfloor
60 .. opcode:: MOV - Move
73 .. opcode:: LIT - Light Coefficients
78 dst.y &= max(src.x, 0) \\
79 dst.z &= (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0 \\
83 .. opcode:: RCP - Reciprocal
85 This instruction replicates its result.
92 .. opcode:: RSQ - Reciprocal Square Root
94 This instruction replicates its result. The results are undefined for src <= 0.
98 dst = \frac{1}{\sqrt{src.x}}
101 .. opcode:: SQRT - Square Root
103 This instruction replicates its result. The results are undefined for src < 0.
110 .. opcode:: EXP - Approximate Exponential Base 2
114 dst.x &= 2^{\lfloor src.x\rfloor} \\
115 dst.y &= src.x - \lfloor src.x\rfloor \\
116 dst.z &= 2^{src.x} \\
120 .. opcode:: LOG - Approximate Logarithm Base 2
124 dst.x &= \lfloor\log_2{|src.x|}\rfloor \\
125 dst.y &= \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}} \\
126 dst.z &= \log_2{|src.x|} \\
130 .. opcode:: MUL - Multiply
134 dst.x = src0.x \times src1.x
136 dst.y = src0.y \times src1.y
138 dst.z = src0.z \times src1.z
140 dst.w = src0.w \times src1.w
143 .. opcode:: ADD - Add
147 dst.x = src0.x + src1.x
149 dst.y = src0.y + src1.y
151 dst.z = src0.z + src1.z
153 dst.w = src0.w + src1.w
156 .. opcode:: DP3 - 3-component Dot Product
158 This instruction replicates its result.
162 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z
165 .. opcode:: DP4 - 4-component Dot Product
167 This instruction replicates its result.
171 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src0.w \times src1.w
174 .. opcode:: DST - Distance Vector
179 dst.y &= src0.y \times src1.y\\
184 .. opcode:: MIN - Minimum
188 dst.x = min(src0.x, src1.x)
190 dst.y = min(src0.y, src1.y)
192 dst.z = min(src0.z, src1.z)
194 dst.w = min(src0.w, src1.w)
197 .. opcode:: MAX - Maximum
201 dst.x = max(src0.x, src1.x)
203 dst.y = max(src0.y, src1.y)
205 dst.z = max(src0.z, src1.z)
207 dst.w = max(src0.w, src1.w)
210 .. opcode:: SLT - Set On Less Than
214 dst.x = (src0.x < src1.x) ? 1.0F : 0.0F
216 dst.y = (src0.y < src1.y) ? 1.0F : 0.0F
218 dst.z = (src0.z < src1.z) ? 1.0F : 0.0F
220 dst.w = (src0.w < src1.w) ? 1.0F : 0.0F
223 .. opcode:: SGE - Set On Greater Equal Than
227 dst.x = (src0.x >= src1.x) ? 1.0F : 0.0F
229 dst.y = (src0.y >= src1.y) ? 1.0F : 0.0F
231 dst.z = (src0.z >= src1.z) ? 1.0F : 0.0F
233 dst.w = (src0.w >= src1.w) ? 1.0F : 0.0F
236 .. opcode:: MAD - Multiply And Add
240 dst.x = src0.x \times src1.x + src2.x
242 dst.y = src0.y \times src1.y + src2.y
244 dst.z = src0.z \times src1.z + src2.z
246 dst.w = src0.w \times src1.w + src2.w
249 .. opcode:: SUB - Subtract
253 dst.x = src0.x - src1.x
255 dst.y = src0.y - src1.y
257 dst.z = src0.z - src1.z
259 dst.w = src0.w - src1.w
262 .. opcode:: LRP - Linear Interpolate
266 dst.x = src0.x \times src1.x + (1 - src0.x) \times src2.x
268 dst.y = src0.y \times src1.y + (1 - src0.y) \times src2.y
270 dst.z = src0.z \times src1.z + (1 - src0.z) \times src2.z
272 dst.w = src0.w \times src1.w + (1 - src0.w) \times src2.w
275 .. opcode:: CND - Condition
279 dst.x = (src2.x > 0.5) ? src0.x : src1.x
281 dst.y = (src2.y > 0.5) ? src0.y : src1.y
283 dst.z = (src2.z > 0.5) ? src0.z : src1.z
285 dst.w = (src2.w > 0.5) ? src0.w : src1.w
288 .. opcode:: DP2A - 2-component Dot Product And Add
292 dst.x = src0.x \times src1.x + src0.y \times src1.y + src2.x
294 dst.y = src0.x \times src1.x + src0.y \times src1.y + src2.x
296 dst.z = src0.x \times src1.x + src0.y \times src1.y + src2.x
298 dst.w = src0.x \times src1.x + src0.y \times src1.y + src2.x
301 .. opcode:: FRC - Fraction
305 dst.x = src.x - \lfloor src.x\rfloor
307 dst.y = src.y - \lfloor src.y\rfloor
309 dst.z = src.z - \lfloor src.z\rfloor
311 dst.w = src.w - \lfloor src.w\rfloor
314 .. opcode:: CLAMP - Clamp
318 dst.x = clamp(src0.x, src1.x, src2.x)
320 dst.y = clamp(src0.y, src1.y, src2.y)
322 dst.z = clamp(src0.z, src1.z, src2.z)
324 dst.w = clamp(src0.w, src1.w, src2.w)
327 .. opcode:: FLR - Floor
329 This is identical to :opcode:`ARL`.
333 dst.x = \lfloor src.x\rfloor
335 dst.y = \lfloor src.y\rfloor
337 dst.z = \lfloor src.z\rfloor
339 dst.w = \lfloor src.w\rfloor
342 .. opcode:: ROUND - Round
355 .. opcode:: EX2 - Exponential Base 2
357 This instruction replicates its result.
364 .. opcode:: LG2 - Logarithm Base 2
366 This instruction replicates its result.
373 .. opcode:: POW - Power
375 This instruction replicates its result.
379 dst = src0.x^{src1.x}
381 .. opcode:: XPD - Cross Product
385 dst.x = src0.y \times src1.z - src1.y \times src0.z
387 dst.y = src0.z \times src1.x - src1.z \times src0.x
389 dst.z = src0.x \times src1.y - src1.x \times src0.y
394 .. opcode:: ABS - Absolute
407 .. opcode:: RCC - Reciprocal Clamped
409 This instruction replicates its result.
411 XXX cleanup on aisle three
415 dst = (1 / src.x) > 0 ? clamp(1 / src.x, 5.42101e-020, 1.84467e+019) : clamp(1 / src.x, -1.84467e+019, -5.42101e-020)
418 .. opcode:: DPH - Homogeneous Dot Product
420 This instruction replicates its result.
424 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
427 .. opcode:: COS - Cosine
429 This instruction replicates its result.
436 .. opcode:: DDX - Derivative Relative To X
440 dst.x = partialx(src.x)
442 dst.y = partialx(src.y)
444 dst.z = partialx(src.z)
446 dst.w = partialx(src.w)
449 .. opcode:: DDY - Derivative Relative To Y
453 dst.x = partialy(src.x)
455 dst.y = partialy(src.y)
457 dst.z = partialy(src.z)
459 dst.w = partialy(src.w)
462 .. opcode:: PK2H - Pack Two 16-bit Floats
467 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
472 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
477 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
482 .. opcode:: RFL - Reflection Vector
486 dst.x = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.x - src1.x
488 dst.y = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.y - src1.y
490 dst.z = 2 \times (src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z) / (src0.x \times src0.x + src0.y \times src0.y + src0.z \times src0.z) \times src0.z - src1.z
496 Considered for removal.
499 .. opcode:: SEQ - Set On Equal
503 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
505 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
507 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
509 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
512 .. opcode:: SFL - Set On False
514 This instruction replicates its result.
522 Considered for removal.
525 .. opcode:: SGT - Set On Greater Than
529 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
531 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
533 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
535 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
538 .. opcode:: SIN - Sine
540 This instruction replicates its result.
547 .. opcode:: SLE - Set On Less Equal Than
551 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
553 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
555 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
557 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
560 .. opcode:: SNE - Set On Not Equal
564 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
566 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
568 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
570 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
573 .. opcode:: STR - Set On True
575 This instruction replicates its result.
582 .. opcode:: TEX - Texture Lookup
584 for array textures src0.y contains the slice for 1D,
585 and src0.z contain the slice for 2D.
587 for shadow textures with no arrays, src0.z contains
590 for shadow textures with arrays, src0.z contains
591 the reference value for 1D arrays, and src0.w contains
592 the reference value for 2D arrays.
594 There is no way to pass a bias in the .w value for
595 shadow arrays, and GLSL doesn't allow this.
596 GLSL does allow cube shadows maps to take a bias value,
597 and we have to determine how this will look in TGSI.
605 dst = texture\_sample(unit, coord, bias)
607 .. opcode:: TXD - Texture Lookup with Derivatives
619 dst = texture\_sample\_deriv(unit, coord, bias, ddx, ddy)
622 .. opcode:: TXP - Projective Texture Lookup
626 coord.x = src0.x / src.w
628 coord.y = src0.y / src.w
630 coord.z = src0.z / src.w
636 dst = texture\_sample(unit, coord, bias)
639 .. opcode:: UP2H - Unpack Two 16-Bit Floats
645 Considered for removal.
647 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
653 Considered for removal.
655 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
661 Considered for removal.
663 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
669 Considered for removal.
671 .. opcode:: X2D - 2D Coordinate Transformation
675 dst.x = src0.x + src1.x \times src2.x + src1.y \times src2.y
677 dst.y = src0.y + src1.x \times src2.z + src1.y \times src2.w
679 dst.z = src0.x + src1.x \times src2.x + src1.y \times src2.y
681 dst.w = src0.y + src1.x \times src2.z + src1.y \times src2.w
685 Considered for removal.
688 .. opcode:: ARA - Address Register Add
694 Considered for removal.
696 .. opcode:: ARR - Address Register Load With Round
709 .. opcode:: SSG - Set Sign
713 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
715 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
717 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
719 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
722 .. opcode:: CMP - Compare
726 dst.x = (src0.x < 0) ? src1.x : src2.x
728 dst.y = (src0.y < 0) ? src1.y : src2.y
730 dst.z = (src0.z < 0) ? src1.z : src2.z
732 dst.w = (src0.w < 0) ? src1.w : src2.w
735 .. opcode:: KILL_IF - Conditional Discard
737 Conditional discard. Allowed in fragment shaders only.
741 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
746 .. opcode:: KILL - Discard
748 Unconditional discard. Allowed in fragment shaders only.
751 .. opcode:: SCS - Sine Cosine
764 .. opcode:: TXB - Texture Lookup With Bias
778 dst = texture\_sample(unit, coord, bias)
781 .. opcode:: NRM - 3-component Vector Normalise
785 dst.x = src.x / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
787 dst.y = src.y / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
789 dst.z = src.z / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
794 .. opcode:: DIV - Divide
798 dst.x = \frac{src0.x}{src1.x}
800 dst.y = \frac{src0.y}{src1.y}
802 dst.z = \frac{src0.z}{src1.z}
804 dst.w = \frac{src0.w}{src1.w}
807 .. opcode:: DP2 - 2-component Dot Product
809 This instruction replicates its result.
813 dst = src0.x \times src1.x + src0.y \times src1.y
816 .. opcode:: TXL - Texture Lookup With explicit LOD
830 dst = texture\_sample(unit, coord, lod)
833 .. opcode:: PUSHA - Push Address Register On Stack
842 Considered for cleanup.
846 Considered for removal.
848 .. opcode:: POPA - Pop Address Register From Stack
857 Considered for cleanup.
861 Considered for removal.
864 .. opcode:: BRA - Branch
870 Considered for removal.
873 .. opcode:: CALLNZ - Subroutine Call If Not Zero
879 Considered for cleanup.
883 Considered for removal.
887 ^^^^^^^^^^^^^^^^^^^^^^^^
889 These opcodes are primarily provided for special-use computational shaders.
890 Support for these opcodes indicated by a special pipe capability bit (TBD).
892 XXX doesn't look like most of the opcodes really belong here.
894 .. opcode:: CEIL - Ceiling
898 dst.x = \lceil src.x\rceil
900 dst.y = \lceil src.y\rceil
902 dst.z = \lceil src.z\rceil
904 dst.w = \lceil src.w\rceil
907 .. opcode:: TRUNC - Truncate
920 .. opcode:: MOD - Modulus
924 dst.x = src0.x \bmod src1.x
926 dst.y = src0.y \bmod src1.y
928 dst.z = src0.z \bmod src1.z
930 dst.w = src0.w \bmod src1.w
933 .. opcode:: UARL - Integer Address Register Load
935 Moves the contents of the source register, assumed to be an integer, into the
936 destination register, which is assumed to be an address (ADDR) register.
939 .. opcode:: SAD - Sum Of Absolute Differences
943 dst.x = |src0.x - src1.x| + src2.x
945 dst.y = |src0.y - src1.y| + src2.y
947 dst.z = |src0.z - src1.z| + src2.z
949 dst.w = |src0.w - src1.w| + src2.w
952 .. opcode:: TXF - Texel Fetch
954 As per NV_gpu_shader4, extract a single texel from a specified texture
955 image. The source sampler may not be a CUBE or SHADOW. src 0 is a
956 four-component signed integer vector used to identify the single texel
957 accessed. 3 components + level. src 1 is a 3 component constant signed
958 integer vector, with each component only have a range of -8..+8 (hw only
959 seems to deal with this range, interface allows for up to unsigned int).
960 TXF(uint_vec coord, int_vec offset).
963 .. opcode:: TXQ - Texture Size Query
965 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
966 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
967 depth), 1D array (width, layers), 2D array (width, height, layers)
973 dst.x = texture\_width(unit, lod)
975 dst.y = texture\_height(unit, lod)
977 dst.z = texture\_depth(unit, lod)
979 .. opcode:: TG4 - Texture Gather
981 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
982 filtering operation and packs them into a single register. Only works with
983 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
984 addressing modes of the sampler and the top level of any mip pyramid are
985 used. Set W to zero. It behaves like the TEX instruction, but a filtered
986 sample is not generated. The four samples that contribute to filtering are
987 placed into xyzw in clockwise order, starting with the (u,v) texture
988 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
989 where the magnitude of the deltas are half a texel.
991 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
992 depth compares, single component selection, and a non-constant offset. It
993 doesn't allow support for the GL independent offset to get i0,j0. This would
994 require another CAP is hw can do it natively. For now we lower that before
1003 dst = texture\_gather4 (unit, coord, component)
1005 (with SM5 - cube array shadow)
1013 dst = texture\_gather (uint, coord, compare)
1017 ^^^^^^^^^^^^^^^^^^^^^^^^
1018 These opcodes are used for integer operations.
1019 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1022 .. opcode:: I2F - Signed Integer To Float
1024 Rounding is unspecified (round to nearest even suggested).
1028 dst.x = (float) src.x
1030 dst.y = (float) src.y
1032 dst.z = (float) src.z
1034 dst.w = (float) src.w
1037 .. opcode:: U2F - Unsigned Integer To Float
1039 Rounding is unspecified (round to nearest even suggested).
1043 dst.x = (float) src.x
1045 dst.y = (float) src.y
1047 dst.z = (float) src.z
1049 dst.w = (float) src.w
1052 .. opcode:: F2I - Float to Signed Integer
1054 Rounding is towards zero (truncate).
1055 Values outside signed range (including NaNs) produce undefined results.
1068 .. opcode:: F2U - Float to Unsigned Integer
1070 Rounding is towards zero (truncate).
1071 Values outside unsigned range (including NaNs) produce undefined results.
1075 dst.x = (unsigned) src.x
1077 dst.y = (unsigned) src.y
1079 dst.z = (unsigned) src.z
1081 dst.w = (unsigned) src.w
1084 .. opcode:: UADD - Integer Add
1086 This instruction works the same for signed and unsigned integers.
1087 The low 32bit of the result is returned.
1091 dst.x = src0.x + src1.x
1093 dst.y = src0.y + src1.y
1095 dst.z = src0.z + src1.z
1097 dst.w = src0.w + src1.w
1100 .. opcode:: UMAD - Integer Multiply And Add
1102 This instruction works the same for signed and unsigned integers.
1103 The multiplication returns the low 32bit (as does the result itself).
1107 dst.x = src0.x \times src1.x + src2.x
1109 dst.y = src0.y \times src1.y + src2.y
1111 dst.z = src0.z \times src1.z + src2.z
1113 dst.w = src0.w \times src1.w + src2.w
1116 .. opcode:: UMUL - Integer Multiply
1118 This instruction works the same for signed and unsigned integers.
1119 The low 32bit of the result is returned.
1123 dst.x = src0.x \times src1.x
1125 dst.y = src0.y \times src1.y
1127 dst.z = src0.z \times src1.z
1129 dst.w = src0.w \times src1.w
1132 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1134 The high 32bits of the multiplication of 2 signed integers are returned.
1138 dst.x = (src0.x \times src1.x) >> 32
1140 dst.y = (src0.y \times src1.y) >> 32
1142 dst.z = (src0.z \times src1.z) >> 32
1144 dst.w = (src0.w \times src1.w) >> 32
1147 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1149 The high 32bits of the multiplication of 2 unsigned integers are returned.
1153 dst.x = (src0.x \times src1.x) >> 32
1155 dst.y = (src0.y \times src1.y) >> 32
1157 dst.z = (src0.z \times src1.z) >> 32
1159 dst.w = (src0.w \times src1.w) >> 32
1162 .. opcode:: IDIV - Signed Integer Division
1164 TBD: behavior for division by zero.
1168 dst.x = src0.x \ src1.x
1170 dst.y = src0.y \ src1.y
1172 dst.z = src0.z \ src1.z
1174 dst.w = src0.w \ src1.w
1177 .. opcode:: UDIV - Unsigned Integer Division
1179 For division by zero, 0xffffffff is returned.
1183 dst.x = src0.x \ src1.x
1185 dst.y = src0.y \ src1.y
1187 dst.z = src0.z \ src1.z
1189 dst.w = src0.w \ src1.w
1192 .. opcode:: UMOD - Unsigned Integer Remainder
1194 If second arg is zero, 0xffffffff is returned.
1198 dst.x = src0.x \ src1.x
1200 dst.y = src0.y \ src1.y
1202 dst.z = src0.z \ src1.z
1204 dst.w = src0.w \ src1.w
1207 .. opcode:: NOT - Bitwise Not
1220 .. opcode:: AND - Bitwise And
1224 dst.x = src0.x \& src1.x
1226 dst.y = src0.y \& src1.y
1228 dst.z = src0.z \& src1.z
1230 dst.w = src0.w \& src1.w
1233 .. opcode:: OR - Bitwise Or
1237 dst.x = src0.x | src1.x
1239 dst.y = src0.y | src1.y
1241 dst.z = src0.z | src1.z
1243 dst.w = src0.w | src1.w
1246 .. opcode:: XOR - Bitwise Xor
1250 dst.x = src0.x \oplus src1.x
1252 dst.y = src0.y \oplus src1.y
1254 dst.z = src0.z \oplus src1.z
1256 dst.w = src0.w \oplus src1.w
1259 .. opcode:: IMAX - Maximum of Signed Integers
1263 dst.x = max(src0.x, src1.x)
1265 dst.y = max(src0.y, src1.y)
1267 dst.z = max(src0.z, src1.z)
1269 dst.w = max(src0.w, src1.w)
1272 .. opcode:: UMAX - Maximum of Unsigned Integers
1276 dst.x = max(src0.x, src1.x)
1278 dst.y = max(src0.y, src1.y)
1280 dst.z = max(src0.z, src1.z)
1282 dst.w = max(src0.w, src1.w)
1285 .. opcode:: IMIN - Minimum of Signed Integers
1289 dst.x = min(src0.x, src1.x)
1291 dst.y = min(src0.y, src1.y)
1293 dst.z = min(src0.z, src1.z)
1295 dst.w = min(src0.w, src1.w)
1298 .. opcode:: UMIN - Minimum of Unsigned Integers
1302 dst.x = min(src0.x, src1.x)
1304 dst.y = min(src0.y, src1.y)
1306 dst.z = min(src0.z, src1.z)
1308 dst.w = min(src0.w, src1.w)
1311 .. opcode:: SHL - Shift Left
1313 The shift count is masked with 0x1f before the shift is applied.
1317 dst.x = src0.x << (0x1f \& src1.x)
1319 dst.y = src0.y << (0x1f \& src1.y)
1321 dst.z = src0.z << (0x1f \& src1.z)
1323 dst.w = src0.w << (0x1f \& src1.w)
1326 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1328 The shift count is masked with 0x1f before the shift is applied.
1332 dst.x = src0.x >> (0x1f \& src1.x)
1334 dst.y = src0.y >> (0x1f \& src1.y)
1336 dst.z = src0.z >> (0x1f \& src1.z)
1338 dst.w = src0.w >> (0x1f \& src1.w)
1341 .. opcode:: USHR - Logical Shift Right
1343 The shift count is masked with 0x1f before the shift is applied.
1347 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1349 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1351 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1353 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1356 .. opcode:: UCMP - Integer Conditional Move
1360 dst.x = src0.x ? src1.x : src2.x
1362 dst.y = src0.y ? src1.y : src2.y
1364 dst.z = src0.z ? src1.z : src2.z
1366 dst.w = src0.w ? src1.w : src2.w
1370 .. opcode:: ISSG - Integer Set Sign
1374 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1376 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1378 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1380 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1384 .. opcode:: FSLT - Float Set On Less Than (ordered)
1386 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1390 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1392 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1394 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1396 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1399 .. opcode:: ISLT - Signed Integer Set On Less Than
1403 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1405 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1407 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1409 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1412 .. opcode:: USLT - Unsigned Integer Set On Less Than
1416 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1418 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1420 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1422 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1425 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1427 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1431 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1433 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1435 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1437 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1440 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1444 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1446 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1448 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1450 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1453 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1457 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1459 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1461 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1463 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1466 .. opcode:: FSEQ - Float Set On Equal (ordered)
1468 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1472 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1474 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1476 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1478 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1481 .. opcode:: USEQ - Integer Set On Equal
1485 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1487 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1489 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1491 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1494 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1496 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1500 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1502 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1504 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1506 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1509 .. opcode:: USNE - Integer Set On Not Equal
1513 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1515 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1517 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1519 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1522 .. opcode:: INEG - Integer Negate
1537 .. opcode:: IABS - Integer Absolute Value
1551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1553 These opcodes are only supported in geometry shaders; they have no meaning
1554 in any other type of shader.
1556 .. opcode:: EMIT - Emit
1558 Generate a new vertex for the current primitive using the values in the
1562 .. opcode:: ENDPRIM - End Primitive
1564 Complete the current primitive (consisting of the emitted vertices),
1565 and start a new one.
1571 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1572 opcodes is determined by a special capability bit, ``GLSL``.
1573 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1575 .. opcode:: CAL - Subroutine Call
1581 .. opcode:: RET - Subroutine Call Return
1586 .. opcode:: CONT - Continue
1588 Unconditionally moves the point of execution to the instruction after the
1589 last bgnloop. The instruction must appear within a bgnloop/endloop.
1593 Support for CONT is determined by a special capability bit,
1594 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1597 .. opcode:: BGNLOOP - Begin a Loop
1599 Start a loop. Must have a matching endloop.
1602 .. opcode:: BGNSUB - Begin Subroutine
1604 Starts definition of a subroutine. Must have a matching endsub.
1607 .. opcode:: ENDLOOP - End a Loop
1609 End a loop started with bgnloop.
1612 .. opcode:: ENDSUB - End Subroutine
1614 Ends definition of a subroutine.
1617 .. opcode:: NOP - No Operation
1622 .. opcode:: BRK - Break
1624 Unconditionally moves the point of execution to the instruction after the
1625 next endloop or endswitch. The instruction must appear within a loop/endloop
1626 or switch/endswitch.
1629 .. opcode:: BREAKC - Break Conditional
1631 Conditionally moves the point of execution to the instruction after the
1632 next endloop or endswitch. The instruction must appear within a loop/endloop
1633 or switch/endswitch.
1634 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1635 as an integer register.
1639 Considered for removal as it's quite inconsistent wrt other opcodes
1640 (could emulate with UIF/BRK/ENDIF).
1643 .. opcode:: IF - Float If
1645 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1649 where src0.x is interpreted as a floating point register.
1652 .. opcode:: UIF - Bitwise If
1654 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1658 where src0.x is interpreted as an integer register.
1661 .. opcode:: ELSE - Else
1663 Starts an else block, after an IF or UIF statement.
1666 .. opcode:: ENDIF - End If
1668 Ends an IF or UIF block.
1671 .. opcode:: SWITCH - Switch
1673 Starts a C-style switch expression. The switch consists of one or multiple
1674 CASE statements, and at most one DEFAULT statement. Execution of a statement
1675 ends when a BRK is hit, but just like in C falling through to other cases
1676 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1677 just as last statement, and fallthrough is allowed into/from it.
1678 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1684 (some instructions here)
1687 (some instructions here)
1690 (some instructions here)
1695 .. opcode:: CASE - Switch case
1697 This represents a switch case label. The src arg must be an integer immediate.
1700 .. opcode:: DEFAULT - Switch default
1702 This represents the default case in the switch, which is taken if no other
1706 .. opcode:: ENDSWITCH - End of switch
1708 Ends a switch expression.
1711 .. opcode:: NRM4 - 4-component Vector Normalise
1713 This instruction replicates its result.
1717 dst = \frac{src.x}{src.x \times src.x + src.y \times src.y + src.z \times src.z + src.w \times src.w}
1725 The double-precision opcodes reinterpret four-component vectors into
1726 two-component vectors with doubled precision in each component.
1728 Support for these opcodes is XXX undecided. :T
1730 .. opcode:: DADD - Add
1734 dst.xy = src0.xy + src1.xy
1736 dst.zw = src0.zw + src1.zw
1739 .. opcode:: DDIV - Divide
1743 dst.xy = src0.xy / src1.xy
1745 dst.zw = src0.zw / src1.zw
1747 .. opcode:: DSEQ - Set on Equal
1751 dst.xy = src0.xy == src1.xy ? 1.0F : 0.0F
1753 dst.zw = src0.zw == src1.zw ? 1.0F : 0.0F
1755 .. opcode:: DSLT - Set on Less than
1759 dst.xy = src0.xy < src1.xy ? 1.0F : 0.0F
1761 dst.zw = src0.zw < src1.zw ? 1.0F : 0.0F
1763 .. opcode:: DFRAC - Fraction
1767 dst.xy = src.xy - \lfloor src.xy\rfloor
1769 dst.zw = src.zw - \lfloor src.zw\rfloor
1772 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1774 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1775 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1776 :math:`dst1 \times 2^{dst0} = src` .
1780 dst0.xy = exp(src.xy)
1782 dst1.xy = frac(src.xy)
1784 dst0.zw = exp(src.zw)
1786 dst1.zw = frac(src.zw)
1788 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1790 This opcode is the inverse of :opcode:`DFRACEXP`.
1794 dst.xy = src0.xy \times 2^{src1.xy}
1796 dst.zw = src0.zw \times 2^{src1.zw}
1798 .. opcode:: DMIN - Minimum
1802 dst.xy = min(src0.xy, src1.xy)
1804 dst.zw = min(src0.zw, src1.zw)
1806 .. opcode:: DMAX - Maximum
1810 dst.xy = max(src0.xy, src1.xy)
1812 dst.zw = max(src0.zw, src1.zw)
1814 .. opcode:: DMUL - Multiply
1818 dst.xy = src0.xy \times src1.xy
1820 dst.zw = src0.zw \times src1.zw
1823 .. opcode:: DMAD - Multiply And Add
1827 dst.xy = src0.xy \times src1.xy + src2.xy
1829 dst.zw = src0.zw \times src1.zw + src2.zw
1832 .. opcode:: DRCP - Reciprocal
1836 dst.xy = \frac{1}{src.xy}
1838 dst.zw = \frac{1}{src.zw}
1840 .. opcode:: DSQRT - Square Root
1844 dst.xy = \sqrt{src.xy}
1846 dst.zw = \sqrt{src.zw}
1849 .. _samplingopcodes:
1851 Resource Sampling Opcodes
1852 ^^^^^^^^^^^^^^^^^^^^^^^^^
1854 Those opcodes follow very closely semantics of the respective Direct3D
1855 instructions. If in doubt double check Direct3D documentation.
1856 Note that the swizzle on SVIEW (src1) determines texel swizzling
1861 Using provided address, sample data from the specified texture using the
1862 filtering mode identified by the gven sampler. The source data may come from
1863 any resource type other than buffers.
1865 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
1867 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
1869 .. opcode:: SAMPLE_I
1871 Simplified alternative to the SAMPLE instruction. Using the provided
1872 integer address, SAMPLE_I fetches data from the specified sampler view
1873 without any filtering. The source data may come from any resource type
1876 Syntax: ``SAMPLE_I dst, address, sampler_view``
1878 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
1880 The 'address' is specified as unsigned integers. If the 'address' is out of
1881 range [0...(# texels - 1)] the result of the fetch is always 0 in all
1882 components. As such the instruction doesn't honor address wrap modes, in
1883 cases where that behavior is desirable 'SAMPLE' instruction should be used.
1884 address.w always provides an unsigned integer mipmap level. If the value is
1885 out of the range then the instruction always returns 0 in all components.
1886 address.yz are ignored for buffers and 1d textures. address.z is ignored
1887 for 1d texture arrays and 2d textures.
1889 For 1D texture arrays address.y provides the array index (also as unsigned
1890 integer). If the value is out of the range of available array indices
1891 [0... (array size - 1)] then the opcode always returns 0 in all components.
1892 For 2D texture arrays address.z provides the array index, otherwise it
1893 exhibits the same behavior as in the case for 1D texture arrays. The exact
1894 semantics of the source address are presented in the table below:
1896 +---------------------------+----+-----+-----+---------+
1897 | resource type | X | Y | Z | W |
1898 +===========================+====+=====+=====+=========+
1899 | ``PIPE_BUFFER`` | x | | | ignored |
1900 +---------------------------+----+-----+-----+---------+
1901 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
1902 +---------------------------+----+-----+-----+---------+
1903 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
1904 +---------------------------+----+-----+-----+---------+
1905 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
1906 +---------------------------+----+-----+-----+---------+
1907 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
1908 +---------------------------+----+-----+-----+---------+
1909 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
1910 +---------------------------+----+-----+-----+---------+
1911 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
1912 +---------------------------+----+-----+-----+---------+
1913 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
1914 +---------------------------+----+-----+-----+---------+
1916 Where 'mpl' is a mipmap level and 'idx' is the array index.
1918 .. opcode:: SAMPLE_I_MS
1920 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
1922 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
1924 .. opcode:: SAMPLE_B
1926 Just like the SAMPLE instruction with the exception that an additional bias
1927 is applied to the level of detail computed as part of the instruction
1930 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
1932 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
1934 .. opcode:: SAMPLE_C
1936 Similar to the SAMPLE instruction but it performs a comparison filter. The
1937 operands to SAMPLE_C are identical to SAMPLE, except that there is an
1938 additional float32 operand, reference value, which must be a register with
1939 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
1940 current samplers compare_func (in pipe_sampler_state) to compare reference
1941 value against the red component value for the surce resource at each texel
1942 that the currently configured texture filter covers based on the provided
1945 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
1947 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
1949 .. opcode:: SAMPLE_C_LZ
1951 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
1954 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
1956 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
1959 .. opcode:: SAMPLE_D
1961 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
1962 the source address in the x direction and the y direction are provided by
1965 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
1967 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
1969 .. opcode:: SAMPLE_L
1971 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
1972 directly as a scalar value, representing no anisotropy.
1974 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
1976 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
1980 Gathers the four texels to be used in a bi-linear filtering operation and
1981 packs them into a single register. Only works with 2D, 2D array, cubemaps,
1982 and cubemaps arrays. For 2D textures, only the addressing modes of the
1983 sampler and the top level of any mip pyramid are used. Set W to zero. It
1984 behaves like the SAMPLE instruction, but a filtered sample is not
1985 generated. The four samples that contribute to filtering are placed into
1986 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
1987 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
1988 magnitude of the deltas are half a texel.
1991 .. opcode:: SVIEWINFO
1993 Query the dimensions of a given sampler view. dst receives width, height,
1994 depth or array size and number of mipmap levels as int4. The dst can have a
1995 writemask which will specify what info is the caller interested in.
1997 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
1999 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2001 src_mip_level is an unsigned integer scalar. If it's out of range then
2002 returns 0 for width, height and depth/array size but the total number of
2003 mipmap is still returned correctly for the given sampler view. The returned
2004 width, height and depth values are for the mipmap level selected by the
2005 src_mip_level and are in the number of texels. For 1d texture array width
2006 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2007 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2008 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2009 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2010 resinfo allowing swizzling dst values is ignored (due to the interaction
2011 with rcpfloat modifier which requires some swizzle handling in the state
2014 .. opcode:: SAMPLE_POS
2016 Query the position of a given sample. dst receives float4 (x, y, 0, 0)
2017 indicated where the sample is located. If the resource is not a multi-sample
2018 resource and not a render target, the result is 0.
2020 .. opcode:: SAMPLE_INFO
2022 dst receives number of samples in x. If the resource is not a multi-sample
2023 resource and not a render target, the result is 0.
2026 .. _resourceopcodes:
2028 Resource Access Opcodes
2029 ^^^^^^^^^^^^^^^^^^^^^^^
2031 .. opcode:: LOAD - Fetch data from a shader resource
2033 Syntax: ``LOAD dst, resource, address``
2035 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
2037 Using the provided integer address, LOAD fetches data
2038 from the specified buffer or texture without any
2041 The 'address' is specified as a vector of unsigned
2042 integers. If the 'address' is out of range the result
2045 Only the first mipmap level of a resource can be read
2046 from using this instruction.
2048 For 1D or 2D texture arrays, the array index is
2049 provided as an unsigned integer in address.y or
2050 address.z, respectively. address.yz are ignored for
2051 buffers and 1D textures. address.z is ignored for 1D
2052 texture arrays and 2D textures. address.w is always
2055 .. opcode:: STORE - Write data to a shader resource
2057 Syntax: ``STORE resource, address, src``
2059 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2061 Using the provided integer address, STORE writes data
2062 to the specified buffer or texture.
2064 The 'address' is specified as a vector of unsigned
2065 integers. If the 'address' is out of range the result
2068 Only the first mipmap level of a resource can be
2069 written to using this instruction.
2071 For 1D or 2D texture arrays, the array index is
2072 provided as an unsigned integer in address.y or
2073 address.z, respectively. address.yz are ignored for
2074 buffers and 1D textures. address.z is ignored for 1D
2075 texture arrays and 2D textures. address.w is always
2079 .. _threadsyncopcodes:
2081 Inter-thread synchronization opcodes
2082 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2084 These opcodes are intended for communication between threads running
2085 within the same compute grid. For now they're only valid in compute
2088 .. opcode:: MFENCE - Memory fence
2090 Syntax: ``MFENCE resource``
2092 Example: ``MFENCE RES[0]``
2094 This opcode forces strong ordering between any memory access
2095 operations that affect the specified resource. This means that
2096 previous loads and stores (and only those) will be performed and
2097 visible to other threads before the program execution continues.
2100 .. opcode:: LFENCE - Load memory fence
2102 Syntax: ``LFENCE resource``
2104 Example: ``LFENCE RES[0]``
2106 Similar to MFENCE, but it only affects the ordering of memory loads.
2109 .. opcode:: SFENCE - Store memory fence
2111 Syntax: ``SFENCE resource``
2113 Example: ``SFENCE RES[0]``
2115 Similar to MFENCE, but it only affects the ordering of memory stores.
2118 .. opcode:: BARRIER - Thread group barrier
2122 This opcode suspends the execution of the current thread until all
2123 the remaining threads in the working group reach the same point of
2124 the program. Results are unspecified if any of the remaining
2125 threads terminates or never reaches an executed BARRIER instruction.
2133 These opcodes provide atomic variants of some common arithmetic and
2134 logical operations. In this context atomicity means that another
2135 concurrent memory access operation that affects the same memory
2136 location is guaranteed to be performed strictly before or after the
2137 entire execution of the atomic operation.
2139 For the moment they're only valid in compute programs.
2141 .. opcode:: ATOMUADD - Atomic integer addition
2143 Syntax: ``ATOMUADD dst, resource, offset, src``
2145 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2147 The following operation is performed atomically on each component:
2151 dst_i = resource[offset]_i
2153 resource[offset]_i = dst_i + src_i
2156 .. opcode:: ATOMXCHG - Atomic exchange
2158 Syntax: ``ATOMXCHG dst, resource, offset, src``
2160 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2162 The following operation is performed atomically on each component:
2166 dst_i = resource[offset]_i
2168 resource[offset]_i = src_i
2171 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2173 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2175 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2177 The following operation is performed atomically on each component:
2181 dst_i = resource[offset]_i
2183 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2186 .. opcode:: ATOMAND - Atomic bitwise And
2188 Syntax: ``ATOMAND dst, resource, offset, src``
2190 Example: ``ATOMAND TEMP[0], RES[0], TEMP[1], TEMP[2]``
2192 The following operation is performed atomically on each component:
2196 dst_i = resource[offset]_i
2198 resource[offset]_i = dst_i \& src_i
2201 .. opcode:: ATOMOR - Atomic bitwise Or
2203 Syntax: ``ATOMOR dst, resource, offset, src``
2205 Example: ``ATOMOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2207 The following operation is performed atomically on each component:
2211 dst_i = resource[offset]_i
2213 resource[offset]_i = dst_i | src_i
2216 .. opcode:: ATOMXOR - Atomic bitwise Xor
2218 Syntax: ``ATOMXOR dst, resource, offset, src``
2220 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2222 The following operation is performed atomically on each component:
2226 dst_i = resource[offset]_i
2228 resource[offset]_i = dst_i \oplus src_i
2231 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2233 Syntax: ``ATOMUMIN dst, resource, offset, src``
2235 Example: ``ATOMUMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2237 The following operation is performed atomically on each component:
2241 dst_i = resource[offset]_i
2243 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2246 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2248 Syntax: ``ATOMUMAX dst, resource, offset, src``
2250 Example: ``ATOMUMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2252 The following operation is performed atomically on each component:
2256 dst_i = resource[offset]_i
2258 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2261 .. opcode:: ATOMIMIN - Atomic signed minimum
2263 Syntax: ``ATOMIMIN dst, resource, offset, src``
2265 Example: ``ATOMIMIN TEMP[0], RES[0], TEMP[1], TEMP[2]``
2267 The following operation is performed atomically on each component:
2271 dst_i = resource[offset]_i
2273 resource[offset]_i = (dst_i < src_i ? dst_i : src_i)
2276 .. opcode:: ATOMIMAX - Atomic signed maximum
2278 Syntax: ``ATOMIMAX dst, resource, offset, src``
2280 Example: ``ATOMIMAX TEMP[0], RES[0], TEMP[1], TEMP[2]``
2282 The following operation is performed atomically on each component:
2286 dst_i = resource[offset]_i
2288 resource[offset]_i = (dst_i > src_i ? dst_i : src_i)
2292 Explanation of symbols used
2293 ------------------------------
2300 :math:`|x|` Absolute value of `x`.
2302 :math:`\lceil x \rceil` Ceiling of `x`.
2304 clamp(x,y,z) Clamp x between y and z.
2305 (x < y) ? y : (x > z) ? z : x
2307 :math:`\lfloor x\rfloor` Floor of `x`.
2309 :math:`\log_2{x}` Logarithm of `x`, base 2.
2311 max(x,y) Maximum of x and y.
2314 min(x,y) Minimum of x and y.
2317 partialx(x) Derivative of x relative to fragment's X.
2319 partialy(x) Derivative of x relative to fragment's Y.
2321 pop() Pop from stack.
2323 :math:`x^y` `x` to the power `y`.
2325 push(x) Push x on stack.
2329 trunc(x) Truncate x, i.e. drop the fraction bits.
2336 discard Discard fragment.
2340 target Label of target instruction.
2351 Declares a register that is will be referenced as an operand in Instruction
2354 File field contains register file that is being declared and is one
2357 UsageMask field specifies which of the register components can be accessed
2358 and is one of TGSI_WRITEMASK.
2360 The Local flag specifies that a given value isn't intended for
2361 subroutine parameter passing and, as a result, the implementation
2362 isn't required to give any guarantees of it being preserved across
2363 subroutine boundaries. As it's merely a compiler hint, the
2364 implementation is free to ignore it.
2366 If Dimension flag is set to 1, a Declaration Dimension token follows.
2368 If Semantic flag is set to 1, a Declaration Semantic token follows.
2370 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2372 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2374 If Array flag is set to 1, a Declaration Array token follows.
2377 ^^^^^^^^^^^^^^^^^^^^^^^^
2379 Declarations can optional have an ArrayID attribute which can be referred by
2380 indirect addressing operands. An ArrayID of zero is reserved and treaded as
2381 if no ArrayID is specified.
2383 If an indirect addressing operand refers to a specific declaration by using
2384 an ArrayID only the registers in this declaration are guaranteed to be
2385 accessed, accessing any register outside this declaration results in undefined
2386 behavior. Note that for compatibility the effective index is zero-based and
2387 not relative to the specified declaration
2389 If no ArrayID is specified with an indirect addressing operand the whole
2390 register file might be accessed by this operand. This is strongly discouraged
2391 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2393 Declaration Semantic
2394 ^^^^^^^^^^^^^^^^^^^^^^^^
2396 Vertex and fragment shader input and output registers may be labeled
2397 with semantic information consisting of a name and index.
2399 Follows Declaration token if Semantic bit is set.
2401 Since its purpose is to link a shader with other stages of the pipeline,
2402 it is valid to follow only those Declaration tokens that declare a register
2403 either in INPUT or OUTPUT file.
2405 SemanticName field contains the semantic name of the register being declared.
2406 There is no default value.
2408 SemanticIndex is an optional subscript that can be used to distinguish
2409 different register declarations with the same semantic name. The default value
2412 The meanings of the individual semantic names are explained in the following
2415 TGSI_SEMANTIC_POSITION
2416 """"""""""""""""""""""
2418 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2419 output register which contains the homogeneous vertex position in the clip
2420 space coordinate system. After clipping, the X, Y and Z components of the
2421 vertex will be divided by the W value to get normalized device coordinates.
2423 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2424 fragment shader input contains the fragment's window position. The X
2425 component starts at zero and always increases from left to right.
2426 The Y component starts at zero and always increases but Y=0 may either
2427 indicate the top of the window or the bottom depending on the fragment
2428 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2429 The Z coordinate ranges from 0 to 1 to represent depth from the front
2430 to the back of the Z buffer. The W component contains the reciprocol
2431 of the interpolated vertex position W component.
2433 Fragment shaders may also declare an output register with
2434 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2435 the fragment shader to change the fragment's Z position.
2442 For vertex shader outputs or fragment shader inputs/outputs, this
2443 label indicates that the resister contains an R,G,B,A color.
2445 Several shader inputs/outputs may contain colors so the semantic index
2446 is used to distinguish them. For example, color[0] may be the diffuse
2447 color while color[1] may be the specular color.
2449 This label is needed so that the flat/smooth shading can be applied
2450 to the right interpolants during rasterization.
2454 TGSI_SEMANTIC_BCOLOR
2455 """"""""""""""""""""
2457 Back-facing colors are only used for back-facing polygons, and are only valid
2458 in vertex shader outputs. After rasterization, all polygons are front-facing
2459 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2460 so all BCOLORs effectively become regular COLORs in the fragment shader.
2466 Vertex shader inputs and outputs and fragment shader inputs may be
2467 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2468 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2469 to compute a fog blend factor which is used to blend the normal fragment color
2470 with a constant fog color. But fog coord really is just an ordinary vec4
2471 register like regular semantics.
2477 Vertex shader input and output registers may be labeled with
2478 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2479 in the form (S, 0, 0, 1). The point size controls the width or diameter
2480 of points for rasterization. This label cannot be used in fragment
2483 When using this semantic, be sure to set the appropriate state in the
2484 :ref:`rasterizer` first.
2487 TGSI_SEMANTIC_TEXCOORD
2488 """"""""""""""""""""""
2490 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2492 Vertex shader outputs and fragment shader inputs may be labeled with
2493 this semantic to make them replaceable by sprite coordinates via the
2494 sprite_coord_enable state in the :ref:`rasterizer`.
2495 The semantic index permitted with this semantic is limited to <= 7.
2497 If the driver does not support TEXCOORD, sprite coordinate replacement
2498 applies to inputs with the GENERIC semantic instead.
2500 The intended use case for this semantic is gl_TexCoord.
2503 TGSI_SEMANTIC_PCOORD
2504 """"""""""""""""""""
2506 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2508 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2509 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2510 the current primitive is a point and point sprites are enabled. Otherwise,
2511 the contents of the register are undefined.
2513 The intended use case for this semantic is gl_PointCoord.
2516 TGSI_SEMANTIC_GENERIC
2517 """""""""""""""""""""
2519 All vertex/fragment shader inputs/outputs not labeled with any other
2520 semantic label can be considered to be generic attributes. Typical
2521 uses of generic inputs/outputs are texcoords and user-defined values.
2524 TGSI_SEMANTIC_NORMAL
2525 """"""""""""""""""""
2527 Indicates that a vertex shader input is a normal vector. This is
2528 typically only used for legacy graphics APIs.
2534 This label applies to fragment shader inputs only and indicates that
2535 the register contains front/back-face information of the form (F, 0,
2536 0, 1). The first component will be positive when the fragment belongs
2537 to a front-facing polygon, and negative when the fragment belongs to a
2538 back-facing polygon.
2541 TGSI_SEMANTIC_EDGEFLAG
2542 """"""""""""""""""""""
2544 For vertex shaders, this sematic label indicates that an input or
2545 output is a boolean edge flag. The register layout is [F, x, x, x]
2546 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2547 simply copies the edge flag input to the edgeflag output.
2549 Edge flags are used to control which lines or points are actually
2550 drawn when the polygon mode converts triangles/quads/polygons into
2554 TGSI_SEMANTIC_STENCIL
2555 """""""""""""""""""""
2557 For fragment shaders, this semantic label indicates that an output
2558 is a writable stencil reference value. Only the Y component is writable.
2559 This allows the fragment shader to change the fragments stencilref value.
2562 TGSI_SEMANTIC_VIEWPORT_INDEX
2563 """"""""""""""""""""""""""""
2565 For geometry shaders, this semantic label indicates that an output
2566 contains the index of the viewport (and scissor) to use.
2567 Only the X value is used.
2573 For geometry shaders, this semantic label indicates that an output
2574 contains the layer value to use for the color and depth/stencil surfaces.
2575 Only the X value is used. (Also known as rendertarget array index.)
2578 TGSI_SEMANTIC_CULLDIST
2579 """"""""""""""""""""""
2581 Used as distance to plane for performing application-defined culling
2582 of individual primitives against a plane. When components of vertex
2583 elements are given this label, these values are assumed to be a
2584 float32 signed distance to a plane. Primitives will be completely
2585 discarded if the plane distance for all of the vertices in the
2586 primitive are < 0. If a vertex has a cull distance of NaN, that
2587 vertex counts as "out" (as if its < 0);
2588 The limits on both clip and cull distances are bound
2589 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2590 the maximum number of components that can be used to hold the
2591 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2592 which specifies the maximum number of registers which can be
2593 annotated with those semantics.
2596 TGSI_SEMANTIC_CLIPDIST
2597 """"""""""""""""""""""
2599 When components of vertex elements are identified this way, these
2600 values are each assumed to be a float32 signed distance to a plane.
2601 Primitive setup only invokes rasterization on pixels for which
2602 the interpolated plane distances are >= 0. Multiple clip planes
2603 can be implemented simultaneously, by annotating multiple
2604 components of one or more vertex elements with the above specified
2605 semantic. The limits on both clip and cull distances are bound
2606 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2607 the maximum number of components that can be used to hold the
2608 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2609 which specifies the maximum number of registers which can be
2610 annotated with those semantics.
2613 Declaration Interpolate
2614 ^^^^^^^^^^^^^^^^^^^^^^^
2616 This token is only valid for fragment shader INPUT declarations.
2618 The Interpolate field specifes the way input is being interpolated by
2619 the rasteriser and is one of TGSI_INTERPOLATE_*.
2621 The CylindricalWrap bitfield specifies which register components
2622 should be subject to cylindrical wrapping when interpolating by the
2623 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2624 should be interpolated according to cylindrical wrapping rules.
2627 Declaration Sampler View
2628 ^^^^^^^^^^^^^^^^^^^^^^^^
2630 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2632 DCL SVIEW[#], resource, type(s)
2634 Declares a shader input sampler view and assigns it to a SVIEW[#]
2637 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2639 type must be 1 or 4 entries (if specifying on a per-component
2640 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
2643 Declaration Resource
2644 ^^^^^^^^^^^^^^^^^^^^
2646 Follows Declaration token if file is TGSI_FILE_RESOURCE.
2648 DCL RES[#], resource [, WR] [, RAW]
2650 Declares a shader input resource and assigns it to a RES[#]
2653 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
2656 If the RAW keyword is not specified, the texture data will be
2657 subject to conversion, swizzling and scaling as required to yield
2658 the specified data type from the physical data format of the bound
2661 If the RAW keyword is specified, no channel conversion will be
2662 performed: the values read for each of the channels (X,Y,Z,W) will
2663 correspond to consecutive words in the same order and format
2664 they're found in memory. No element-to-address conversion will be
2665 performed either: the value of the provided X coordinate will be
2666 interpreted in byte units instead of texel units. The result of
2667 accessing a misaligned address is undefined.
2669 Usage of the STORE opcode is only allowed if the WR (writable) flag
2674 ^^^^^^^^^^^^^^^^^^^^^^^^
2676 Properties are general directives that apply to the whole TGSI program.
2681 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
2682 The default value is UPPER_LEFT.
2684 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
2685 increase downward and rightward.
2686 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
2687 increase upward and rightward.
2689 OpenGL defaults to LOWER_LEFT, and is configurable with the
2690 GL_ARB_fragment_coord_conventions extension.
2692 DirectX 9/10 use UPPER_LEFT.
2694 FS_COORD_PIXEL_CENTER
2695 """""""""""""""""""""
2697 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
2698 The default value is HALF_INTEGER.
2700 If HALF_INTEGER, the fractionary part of the position will be 0.5
2701 If INTEGER, the fractionary part of the position will be 0.0
2703 Note that this does not affect the set of fragments generated by
2704 rasterization, which is instead controlled by half_pixel_center in the
2707 OpenGL defaults to HALF_INTEGER, and is configurable with the
2708 GL_ARB_fragment_coord_conventions extension.
2710 DirectX 9 uses INTEGER.
2711 DirectX 10 uses HALF_INTEGER.
2713 FS_COLOR0_WRITES_ALL_CBUFS
2714 """"""""""""""""""""""""""
2715 Specifies that writes to the fragment shader color 0 are replicated to all
2716 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
2717 fragData is directed to a single color buffer, but fragColor is broadcast.
2720 """"""""""""""""""""""""""
2721 If this property is set on the program bound to the shader stage before the
2722 fragment shader, user clip planes should have no effect (be disabled) even if
2723 that shader does not write to any clip distance outputs and the rasterizer's
2724 clip_plane_enable is non-zero.
2725 This property is only supported by drivers that also support shader clip
2727 This is useful for APIs that don't have UCPs and where clip distances written
2728 by a shader cannot be disabled.
2731 Texture Sampling and Texture Formats
2732 ------------------------------------
2734 This table shows how texture image components are returned as (x,y,z,w) tuples
2735 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
2736 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
2739 +--------------------+--------------+--------------------+--------------+
2740 | Texture Components | Gallium | OpenGL | Direct3D 9 |
2741 +====================+==============+====================+==============+
2742 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
2743 +--------------------+--------------+--------------------+--------------+
2744 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
2745 +--------------------+--------------+--------------------+--------------+
2746 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
2747 +--------------------+--------------+--------------------+--------------+
2748 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
2749 +--------------------+--------------+--------------------+--------------+
2750 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
2751 +--------------------+--------------+--------------------+--------------+
2752 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
2753 +--------------------+--------------+--------------------+--------------+
2754 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
2755 +--------------------+--------------+--------------------+--------------+
2756 | I | (i, i, i, i) | (i, i, i, i) | N/A |
2757 +--------------------+--------------+--------------------+--------------+
2758 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
2759 | | | [#envmap-bumpmap]_ | |
2760 +--------------------+--------------+--------------------+--------------+
2761 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
2762 | | | [#depth-tex-mode]_ | |
2763 +--------------------+--------------+--------------------+--------------+
2764 | S | (s, s, s, s) | unknown | unknown |
2765 +--------------------+--------------+--------------------+--------------+
2767 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
2768 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
2769 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.