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:`Double Opcodes`.
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
81 dst.z = (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0
86 .. opcode:: RCP - Reciprocal
88 This instruction replicates its result.
95 .. opcode:: RSQ - Reciprocal Square Root
97 This instruction replicates its result. The results are undefined for src <= 0.
101 dst = \frac{1}{\sqrt{src.x}}
104 .. opcode:: SQRT - Square Root
106 This instruction replicates its result. The results are undefined for src < 0.
113 .. opcode:: EXP - Approximate Exponential Base 2
117 dst.x = 2^{\lfloor src.x\rfloor}
119 dst.y = src.x - \lfloor src.x\rfloor
126 .. opcode:: LOG - Approximate Logarithm Base 2
130 dst.x = \lfloor\log_2{|src.x|}\rfloor
132 dst.y = \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}}
134 dst.z = \log_2{|src.x|}
139 .. opcode:: MUL - Multiply
143 dst.x = src0.x \times src1.x
145 dst.y = src0.y \times src1.y
147 dst.z = src0.z \times src1.z
149 dst.w = src0.w \times src1.w
152 .. opcode:: ADD - Add
156 dst.x = src0.x + src1.x
158 dst.y = src0.y + src1.y
160 dst.z = src0.z + src1.z
162 dst.w = src0.w + src1.w
165 .. opcode:: DP3 - 3-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
174 .. opcode:: DP4 - 4-component Dot Product
176 This instruction replicates its result.
180 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src0.w \times src1.w
183 .. opcode:: DST - Distance Vector
189 dst.y = src0.y \times src1.y
196 .. opcode:: MIN - Minimum
200 dst.x = min(src0.x, src1.x)
202 dst.y = min(src0.y, src1.y)
204 dst.z = min(src0.z, src1.z)
206 dst.w = min(src0.w, src1.w)
209 .. opcode:: MAX - Maximum
213 dst.x = max(src0.x, src1.x)
215 dst.y = max(src0.y, src1.y)
217 dst.z = max(src0.z, src1.z)
219 dst.w = max(src0.w, src1.w)
222 .. opcode:: SLT - Set On Less Than
226 dst.x = (src0.x < src1.x) ? 1 : 0
228 dst.y = (src0.y < src1.y) ? 1 : 0
230 dst.z = (src0.z < src1.z) ? 1 : 0
232 dst.w = (src0.w < src1.w) ? 1 : 0
235 .. opcode:: SGE - Set On Greater Equal Than
239 dst.x = (src0.x >= src1.x) ? 1 : 0
241 dst.y = (src0.y >= src1.y) ? 1 : 0
243 dst.z = (src0.z >= src1.z) ? 1 : 0
245 dst.w = (src0.w >= src1.w) ? 1 : 0
248 .. opcode:: MAD - Multiply And Add
252 dst.x = src0.x \times src1.x + src2.x
254 dst.y = src0.y \times src1.y + src2.y
256 dst.z = src0.z \times src1.z + src2.z
258 dst.w = src0.w \times src1.w + src2.w
261 .. opcode:: SUB - Subtract
265 dst.x = src0.x - src1.x
267 dst.y = src0.y - src1.y
269 dst.z = src0.z - src1.z
271 dst.w = src0.w - src1.w
274 .. opcode:: LRP - Linear Interpolate
278 dst.x = src0.x \times src1.x + (1 - src0.x) \times src2.x
280 dst.y = src0.y \times src1.y + (1 - src0.y) \times src2.y
282 dst.z = src0.z \times src1.z + (1 - src0.z) \times src2.z
284 dst.w = src0.w \times src1.w + (1 - src0.w) \times src2.w
287 .. opcode:: CND - Condition
291 dst.x = (src2.x > 0.5) ? src0.x : src1.x
293 dst.y = (src2.y > 0.5) ? src0.y : src1.y
295 dst.z = (src2.z > 0.5) ? src0.z : src1.z
297 dst.w = (src2.w > 0.5) ? src0.w : src1.w
300 .. opcode:: DP2A - 2-component Dot Product And Add
304 dst.x = src0.x \times src1.x + src0.y \times src1.y + src2.x
306 dst.y = src0.x \times src1.x + src0.y \times src1.y + src2.x
308 dst.z = src0.x \times src1.x + src0.y \times src1.y + src2.x
310 dst.w = src0.x \times src1.x + src0.y \times src1.y + src2.x
313 .. opcode:: FRC - Fraction
317 dst.x = src.x - \lfloor src.x\rfloor
319 dst.y = src.y - \lfloor src.y\rfloor
321 dst.z = src.z - \lfloor src.z\rfloor
323 dst.w = src.w - \lfloor src.w\rfloor
326 .. opcode:: CLAMP - Clamp
330 dst.x = clamp(src0.x, src1.x, src2.x)
332 dst.y = clamp(src0.y, src1.y, src2.y)
334 dst.z = clamp(src0.z, src1.z, src2.z)
336 dst.w = clamp(src0.w, src1.w, src2.w)
339 .. opcode:: FLR - Floor
341 This is identical to :opcode:`ARL`.
345 dst.x = \lfloor src.x\rfloor
347 dst.y = \lfloor src.y\rfloor
349 dst.z = \lfloor src.z\rfloor
351 dst.w = \lfloor src.w\rfloor
354 .. opcode:: ROUND - Round
367 .. opcode:: EX2 - Exponential Base 2
369 This instruction replicates its result.
376 .. opcode:: LG2 - Logarithm Base 2
378 This instruction replicates its result.
385 .. opcode:: POW - Power
387 This instruction replicates its result.
391 dst = src0.x^{src1.x}
393 .. opcode:: XPD - Cross Product
397 dst.x = src0.y \times src1.z - src1.y \times src0.z
399 dst.y = src0.z \times src1.x - src1.z \times src0.x
401 dst.z = src0.x \times src1.y - src1.x \times src0.y
406 .. opcode:: ABS - Absolute
419 .. opcode:: RCC - Reciprocal Clamped
421 This instruction replicates its result.
423 XXX cleanup on aisle three
427 dst = (1 / src.x) > 0 ? clamp(1 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1 / src.x, -1.884467e+019, -5.42101e-020)
430 .. opcode:: DPH - Homogeneous Dot Product
432 This instruction replicates its result.
436 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
439 .. opcode:: COS - Cosine
441 This instruction replicates its result.
448 .. opcode:: DDX - Derivative Relative To X
452 dst.x = partialx(src.x)
454 dst.y = partialx(src.y)
456 dst.z = partialx(src.z)
458 dst.w = partialx(src.w)
461 .. opcode:: DDY - Derivative Relative To Y
465 dst.x = partialy(src.x)
467 dst.y = partialy(src.y)
469 dst.z = partialy(src.z)
471 dst.w = partialy(src.w)
474 .. opcode:: PK2H - Pack Two 16-bit Floats
479 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
484 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
489 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
494 .. opcode:: RFL - Reflection Vector
498 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
500 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
502 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
508 Considered for removal.
511 .. opcode:: SEQ - Set On Equal
515 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
517 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
519 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
521 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
524 .. opcode:: SFL - Set On False
526 This instruction replicates its result.
534 Considered for removal.
537 .. opcode:: SGT - Set On Greater Than
541 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
543 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
545 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
547 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
550 .. opcode:: SIN - Sine
552 This instruction replicates its result.
559 .. opcode:: SLE - Set On Less Equal Than
563 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
565 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
567 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
569 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
572 .. opcode:: SNE - Set On Not Equal
576 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
578 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
580 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
582 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
585 .. opcode:: STR - Set On True
587 This instruction replicates its result.
594 .. opcode:: TEX - Texture Lookup
602 dst = texture_sample(unit, coord, bias)
604 for array textures src0.y contains the slice for 1D,
605 and src0.z contain the slice for 2D.
606 for shadow textures with no arrays, src0.z contains
608 for shadow textures with arrays, src0.z contains
609 the reference value for 1D arrays, and src0.w contains
610 the reference value for 2D arrays.
611 There is no way to pass a bias in the .w value for
612 shadow arrays, and GLSL doesn't allow this.
613 GLSL does allow cube shadows maps to take a bias value,
614 and we have to determine how this will look in TGSI.
616 .. opcode:: TXD - Texture Lookup with Derivatives
628 dst = texture_sample_deriv(unit, coord, bias, ddx, ddy)
631 .. opcode:: TXP - Projective Texture Lookup
635 coord.x = src0.x / src.w
637 coord.y = src0.y / src.w
639 coord.z = src0.z / src.w
645 dst = texture_sample(unit, coord, bias)
648 .. opcode:: UP2H - Unpack Two 16-Bit Floats
654 Considered for removal.
656 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
662 Considered for removal.
664 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
670 Considered for removal.
672 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
678 Considered for removal.
680 .. opcode:: X2D - 2D Coordinate Transformation
684 dst.x = src0.x + src1.x \times src2.x + src1.y \times src2.y
686 dst.y = src0.y + src1.x \times src2.z + src1.y \times src2.w
688 dst.z = src0.x + src1.x \times src2.x + src1.y \times src2.y
690 dst.w = src0.y + src1.x \times src2.z + src1.y \times src2.w
694 Considered for removal.
697 .. opcode:: ARA - Address Register Add
703 Considered for removal.
705 .. opcode:: ARR - Address Register Load With Round
718 .. opcode:: SSG - Set Sign
722 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
724 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
726 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
728 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
731 .. opcode:: CMP - Compare
735 dst.x = (src0.x < 0) ? src1.x : src2.x
737 dst.y = (src0.y < 0) ? src1.y : src2.y
739 dst.z = (src0.z < 0) ? src1.z : src2.z
741 dst.w = (src0.w < 0) ? src1.w : src2.w
744 .. opcode:: KILL_IF - Conditional Discard
746 Conditional discard. Allowed in fragment shaders only.
750 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
755 .. opcode:: KILL - Discard
757 Unconditional discard. Allowed in fragment shaders only.
760 .. opcode:: SCS - Sine Cosine
773 .. opcode:: TXB - Texture Lookup With Bias
787 dst = texture_sample(unit, coord, bias)
790 .. opcode:: NRM - 3-component Vector Normalise
794 dst.x = src.x / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
796 dst.y = src.y / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
798 dst.z = src.z / (src.x \times src.x + src.y \times src.y + src.z \times src.z)
803 .. opcode:: DIV - Divide
807 dst.x = \frac{src0.x}{src1.x}
809 dst.y = \frac{src0.y}{src1.y}
811 dst.z = \frac{src0.z}{src1.z}
813 dst.w = \frac{src0.w}{src1.w}
816 .. opcode:: DP2 - 2-component Dot Product
818 This instruction replicates its result.
822 dst = src0.x \times src1.x + src0.y \times src1.y
825 .. opcode:: TXL - Texture Lookup With explicit LOD
839 dst = texture_sample(unit, coord, lod)
842 .. opcode:: PUSHA - Push Address Register On Stack
851 Considered for cleanup.
855 Considered for removal.
857 .. opcode:: POPA - Pop Address Register From Stack
866 Considered for cleanup.
870 Considered for removal.
873 .. opcode:: BRA - Branch
879 Considered for removal.
882 .. opcode:: CALLNZ - Subroutine Call If Not Zero
888 Considered for cleanup.
892 Considered for removal.
896 ^^^^^^^^^^^^^^^^^^^^^^^^
898 These opcodes are primarily provided for special-use computational shaders.
899 Support for these opcodes indicated by a special pipe capability bit (TBD).
901 XXX doesn't look like most of the opcodes really belong here.
903 .. opcode:: CEIL - Ceiling
907 dst.x = \lceil src.x\rceil
909 dst.y = \lceil src.y\rceil
911 dst.z = \lceil src.z\rceil
913 dst.w = \lceil src.w\rceil
916 .. opcode:: TRUNC - Truncate
929 .. opcode:: MOD - Modulus
933 dst.x = src0.x \bmod src1.x
935 dst.y = src0.y \bmod src1.y
937 dst.z = src0.z \bmod src1.z
939 dst.w = src0.w \bmod src1.w
942 .. opcode:: UARL - Integer Address Register Load
944 Moves the contents of the source register, assumed to be an integer, into the
945 destination register, which is assumed to be an address (ADDR) register.
948 .. opcode:: SAD - Sum Of Absolute Differences
952 dst.x = |src0.x - src1.x| + src2.x
954 dst.y = |src0.y - src1.y| + src2.y
956 dst.z = |src0.z - src1.z| + src2.z
958 dst.w = |src0.w - src1.w| + src2.w
961 .. opcode:: TXF - Texel Fetch (as per NV_gpu_shader4), extract a single texel
962 from a specified texture image. The source sampler may
963 not be a CUBE or SHADOW.
964 src 0 is a four-component signed integer vector used to
965 identify the single texel accessed. 3 components + level.
966 src 1 is a 3 component constant signed integer vector,
967 with each component only have a range of
968 -8..+8 (hw only seems to deal with this range, interface
969 allows for up to unsigned int).
970 TXF(uint_vec coord, int_vec offset).
973 .. opcode:: TXQ - Texture Size Query (as per NV_gpu_program4)
974 retrieve the dimensions of the texture
975 depending on the target. For 1D (width), 2D/RECT/CUBE
976 (width, height), 3D (width, height, depth),
977 1D array (width, layers), 2D array (width, height, layers)
983 dst.x = texture_width(unit, lod)
985 dst.y = texture_height(unit, lod)
987 dst.z = texture_depth(unit, lod)
991 ^^^^^^^^^^^^^^^^^^^^^^^^
992 These opcodes are used for integer operations.
993 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
996 .. opcode:: I2F - Signed Integer To Float
998 Rounding is unspecified (round to nearest even suggested).
1002 dst.x = (float) src.x
1004 dst.y = (float) src.y
1006 dst.z = (float) src.z
1008 dst.w = (float) src.w
1011 .. opcode:: U2F - Unsigned Integer To Float
1013 Rounding is unspecified (round to nearest even suggested).
1017 dst.x = (float) src.x
1019 dst.y = (float) src.y
1021 dst.z = (float) src.z
1023 dst.w = (float) src.w
1026 .. opcode:: F2I - Float to Signed Integer
1028 Rounding is towards zero (truncate).
1029 Values outside signed range (including NaNs) produce undefined results.
1042 .. opcode:: F2U - Float to Unsigned Integer
1044 Rounding is towards zero (truncate).
1045 Values outside unsigned range (including NaNs) produce undefined results.
1049 dst.x = (unsigned) src.x
1051 dst.y = (unsigned) src.y
1053 dst.z = (unsigned) src.z
1055 dst.w = (unsigned) src.w
1058 .. opcode:: UADD - Integer Add
1060 This instruction works the same for signed and unsigned integers.
1061 The low 32bit of the result is returned.
1065 dst.x = src0.x + src1.x
1067 dst.y = src0.y + src1.y
1069 dst.z = src0.z + src1.z
1071 dst.w = src0.w + src1.w
1074 .. opcode:: UMAD - Integer Multiply And Add
1076 This instruction works the same for signed and unsigned integers.
1077 The multiplication returns the low 32bit (as does the result itself).
1081 dst.x = src0.x \times src1.x + src2.x
1083 dst.y = src0.y \times src1.y + src2.y
1085 dst.z = src0.z \times src1.z + src2.z
1087 dst.w = src0.w \times src1.w + src2.w
1090 .. opcode:: UMUL - Integer Multiply
1092 This instruction works the same for signed and unsigned integers.
1093 The low 32bit of the result is returned.
1097 dst.x = src0.x \times src1.x
1099 dst.y = src0.y \times src1.y
1101 dst.z = src0.z \times src1.z
1103 dst.w = src0.w \times src1.w
1106 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1108 The high 32bits of the multiplication of 2 signed integers are returned.
1112 dst.x = (src0.x \times src1.x) >> 32
1114 dst.y = (src0.y \times src1.y) >> 32
1116 dst.z = (src0.z \times src1.z) >> 32
1118 dst.w = (src0.w \times src1.w) >> 32
1121 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1123 The high 32bits of the multiplication of 2 unsigned integers are returned.
1127 dst.x = (src0.x \times src1.x) >> 32
1129 dst.y = (src0.y \times src1.y) >> 32
1131 dst.z = (src0.z \times src1.z) >> 32
1133 dst.w = (src0.w \times src1.w) >> 32
1136 .. opcode:: IDIV - Signed Integer Division
1138 TBD: behavior for division by zero.
1142 dst.x = src0.x \ src1.x
1144 dst.y = src0.y \ src1.y
1146 dst.z = src0.z \ src1.z
1148 dst.w = src0.w \ src1.w
1151 .. opcode:: UDIV - Unsigned Integer Division
1153 For division by zero, 0xffffffff is returned.
1157 dst.x = src0.x \ src1.x
1159 dst.y = src0.y \ src1.y
1161 dst.z = src0.z \ src1.z
1163 dst.w = src0.w \ src1.w
1166 .. opcode:: UMOD - Unsigned Integer Remainder
1168 If second arg is zero, 0xffffffff is returned.
1172 dst.x = src0.x \ src1.x
1174 dst.y = src0.y \ src1.y
1176 dst.z = src0.z \ src1.z
1178 dst.w = src0.w \ src1.w
1181 .. opcode:: NOT - Bitwise Not
1194 .. opcode:: AND - Bitwise And
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:: OR - Bitwise Or
1211 dst.x = src0.x | src1.x
1213 dst.y = src0.y | src1.y
1215 dst.z = src0.z | src1.z
1217 dst.w = src0.w | src1.w
1220 .. opcode:: XOR - Bitwise Xor
1224 dst.x = src0.x \oplus src1.x
1226 dst.y = src0.y \oplus src1.y
1228 dst.z = src0.z \oplus src1.z
1230 dst.w = src0.w \oplus src1.w
1233 .. opcode:: IMAX - Maximum of Signed Integers
1237 dst.x = max(src0.x, src1.x)
1239 dst.y = max(src0.y, src1.y)
1241 dst.z = max(src0.z, src1.z)
1243 dst.w = max(src0.w, src1.w)
1246 .. opcode:: UMAX - Maximum of Unsigned Integers
1250 dst.x = max(src0.x, src1.x)
1252 dst.y = max(src0.y, src1.y)
1254 dst.z = max(src0.z, src1.z)
1256 dst.w = max(src0.w, src1.w)
1259 .. opcode:: IMIN - Minimum of Signed Integers
1263 dst.x = min(src0.x, src1.x)
1265 dst.y = min(src0.y, src1.y)
1267 dst.z = min(src0.z, src1.z)
1269 dst.w = min(src0.w, src1.w)
1272 .. opcode:: UMIN - Minimum of Unsigned Integers
1276 dst.x = min(src0.x, src1.x)
1278 dst.y = min(src0.y, src1.y)
1280 dst.z = min(src0.z, src1.z)
1282 dst.w = min(src0.w, src1.w)
1285 .. opcode:: SHL - Shift Left
1287 The shift count is masked with 0x1f before the shift is applied.
1291 dst.x = src0.x << (0x1f & src1.x)
1293 dst.y = src0.y << (0x1f & src1.y)
1295 dst.z = src0.z << (0x1f & src1.z)
1297 dst.w = src0.w << (0x1f & src1.w)
1300 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1302 The shift count is masked with 0x1f before the shift is applied.
1306 dst.x = src0.x >> (0x1f & src1.x)
1308 dst.y = src0.y >> (0x1f & src1.y)
1310 dst.z = src0.z >> (0x1f & src1.z)
1312 dst.w = src0.w >> (0x1f & src1.w)
1315 .. opcode:: USHR - Logical Shift Right
1317 The shift count is masked with 0x1f before the shift is applied.
1321 dst.x = src0.x >> (unsigned) (0x1f & src1.x)
1323 dst.y = src0.y >> (unsigned) (0x1f & src1.y)
1325 dst.z = src0.z >> (unsigned) (0x1f & src1.z)
1327 dst.w = src0.w >> (unsigned) (0x1f & src1.w)
1330 .. opcode:: UCMP - Integer Conditional Move
1334 dst.x = src0.x ? src1.x : src2.x
1336 dst.y = src0.y ? src1.y : src2.y
1338 dst.z = src0.z ? src1.z : src2.z
1340 dst.w = src0.w ? src1.w : src2.w
1344 .. opcode:: ISSG - Integer Set Sign
1348 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1350 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1352 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1354 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1358 .. opcode:: FSLT - Float Set On Less Than (ordered)
1360 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1364 dst.x = (src0.x < src1.x) ? ~0 : 0
1366 dst.y = (src0.y < src1.y) ? ~0 : 0
1368 dst.z = (src0.z < src1.z) ? ~0 : 0
1370 dst.w = (src0.w < src1.w) ? ~0 : 0
1373 .. opcode:: ISLT - Signed Integer Set On Less Than
1377 dst.x = (src0.x < src1.x) ? ~0 : 0
1379 dst.y = (src0.y < src1.y) ? ~0 : 0
1381 dst.z = (src0.z < src1.z) ? ~0 : 0
1383 dst.w = (src0.w < src1.w) ? ~0 : 0
1386 .. opcode:: USLT - Unsigned Integer Set On Less Than
1390 dst.x = (src0.x < src1.x) ? ~0 : 0
1392 dst.y = (src0.y < src1.y) ? ~0 : 0
1394 dst.z = (src0.z < src1.z) ? ~0 : 0
1396 dst.w = (src0.w < src1.w) ? ~0 : 0
1399 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1401 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1405 dst.x = (src0.x >= src1.x) ? ~0 : 0
1407 dst.y = (src0.y >= src1.y) ? ~0 : 0
1409 dst.z = (src0.z >= src1.z) ? ~0 : 0
1411 dst.w = (src0.w >= src1.w) ? ~0 : 0
1414 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1418 dst.x = (src0.x >= src1.x) ? ~0 : 0
1420 dst.y = (src0.y >= src1.y) ? ~0 : 0
1422 dst.z = (src0.z >= src1.z) ? ~0 : 0
1424 dst.w = (src0.w >= src1.w) ? ~0 : 0
1427 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1431 dst.x = (src0.x >= src1.x) ? ~0 : 0
1433 dst.y = (src0.y >= src1.y) ? ~0 : 0
1435 dst.z = (src0.z >= src1.z) ? ~0 : 0
1437 dst.w = (src0.w >= src1.w) ? ~0 : 0
1440 .. opcode:: FSEQ - Float Set On Equal (ordered)
1442 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1446 dst.x = (src0.x == src1.x) ? ~0 : 0
1448 dst.y = (src0.y == src1.y) ? ~0 : 0
1450 dst.z = (src0.z == src1.z) ? ~0 : 0
1452 dst.w = (src0.w == src1.w) ? ~0 : 0
1455 .. opcode:: USEQ - Integer Set On Equal
1459 dst.x = (src0.x == src1.x) ? ~0 : 0
1461 dst.y = (src0.y == src1.y) ? ~0 : 0
1463 dst.z = (src0.z == src1.z) ? ~0 : 0
1465 dst.w = (src0.w == src1.w) ? ~0 : 0
1468 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1470 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1474 dst.x = (src0.x != src1.x) ? ~0 : 0
1476 dst.y = (src0.y != src1.y) ? ~0 : 0
1478 dst.z = (src0.z != src1.z) ? ~0 : 0
1480 dst.w = (src0.w != src1.w) ? ~0 : 0
1483 .. opcode:: USNE - Integer Set On Not Equal
1487 dst.x = (src0.x != src1.x) ? ~0 : 0
1489 dst.y = (src0.y != src1.y) ? ~0 : 0
1491 dst.z = (src0.z != src1.z) ? ~0 : 0
1493 dst.w = (src0.w != src1.w) ? ~0 : 0
1496 .. opcode:: INEG - Integer Negate
1511 .. opcode:: IABS - Integer Absolute Value
1525 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1527 These opcodes are only supported in geometry shaders; they have no meaning
1528 in any other type of shader.
1530 .. opcode:: EMIT - Emit
1532 Generate a new vertex for the current primitive using the values in the
1536 .. opcode:: ENDPRIM - End Primitive
1538 Complete the current primitive (consisting of the emitted vertices),
1539 and start a new one.
1545 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1546 opcodes is determined by a special capability bit, ``GLSL``.
1547 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1549 .. opcode:: CAL - Subroutine Call
1555 .. opcode:: RET - Subroutine Call Return
1560 .. opcode:: CONT - Continue
1562 Unconditionally moves the point of execution to the instruction after the
1563 last bgnloop. The instruction must appear within a bgnloop/endloop.
1567 Support for CONT is determined by a special capability bit,
1568 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1571 .. opcode:: BGNLOOP - Begin a Loop
1573 Start a loop. Must have a matching endloop.
1576 .. opcode:: BGNSUB - Begin Subroutine
1578 Starts definition of a subroutine. Must have a matching endsub.
1581 .. opcode:: ENDLOOP - End a Loop
1583 End a loop started with bgnloop.
1586 .. opcode:: ENDSUB - End Subroutine
1588 Ends definition of a subroutine.
1591 .. opcode:: NOP - No Operation
1596 .. opcode:: BRK - Break
1598 Unconditionally moves the point of execution to the instruction after the
1599 next endloop or endswitch. The instruction must appear within a loop/endloop
1600 or switch/endswitch.
1603 .. opcode:: BREAKC - Break Conditional
1605 Conditionally moves the point of execution to the instruction after the
1606 next endloop or endswitch. The instruction must appear within a loop/endloop
1607 or switch/endswitch.
1608 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1609 as an integer register.
1613 Considered for removal as it's quite inconsistent wrt other opcodes
1614 (could emulate with UIF/BRK/ENDIF).
1617 .. opcode:: IF - Float If
1619 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1623 where src0.x is interpreted as a floating point register.
1626 .. opcode:: UIF - Bitwise If
1628 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1632 where src0.x is interpreted as an integer register.
1635 .. opcode:: ELSE - Else
1637 Starts an else block, after an IF or UIF statement.
1640 .. opcode:: ENDIF - End If
1642 Ends an IF or UIF block.
1645 .. opcode:: SWITCH - Switch
1647 Starts a C-style switch expression. The switch consists of one or multiple
1648 CASE statements, and at most one DEFAULT statement. Execution of a statement
1649 ends when a BRK is hit, but just like in C falling through to other cases
1650 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1651 just as last statement, and fallthrough is allowed into/from it.
1652 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1657 (some instructions here)
1660 (some instructions here)
1663 (some instructions here)
1668 .. opcode:: CASE - Switch case
1670 This represents a switch case label. The src arg must be an integer immediate.
1673 .. opcode:: DEFAULT - Switch default
1675 This represents the default case in the switch, which is taken if no other
1679 .. opcode:: ENDSWITCH - End of switch
1681 Ends a switch expression.
1684 .. opcode:: NRM4 - 4-component Vector Normalise
1686 This instruction replicates its result.
1690 dst = \frac{src.x}{src.x \times src.x + src.y \times src.y + src.z \times src.z + src.w \times src.w}
1698 The double-precision opcodes reinterpret four-component vectors into
1699 two-component vectors with doubled precision in each component.
1701 Support for these opcodes is XXX undecided. :T
1703 .. opcode:: DADD - Add
1707 dst.xy = src0.xy + src1.xy
1709 dst.zw = src0.zw + src1.zw
1712 .. opcode:: DDIV - Divide
1716 dst.xy = src0.xy / src1.xy
1718 dst.zw = src0.zw / src1.zw
1720 .. opcode:: DSEQ - Set on Equal
1724 dst.xy = src0.xy == src1.xy ? 1.0F : 0.0F
1726 dst.zw = src0.zw == src1.zw ? 1.0F : 0.0F
1728 .. opcode:: DSLT - Set on Less than
1732 dst.xy = src0.xy < src1.xy ? 1.0F : 0.0F
1734 dst.zw = src0.zw < src1.zw ? 1.0F : 0.0F
1736 .. opcode:: DFRAC - Fraction
1740 dst.xy = src.xy - \lfloor src.xy\rfloor
1742 dst.zw = src.zw - \lfloor src.zw\rfloor
1745 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1747 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1748 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1749 :math:`dst1 \times 2^{dst0} = src` .
1753 dst0.xy = exp(src.xy)
1755 dst1.xy = frac(src.xy)
1757 dst0.zw = exp(src.zw)
1759 dst1.zw = frac(src.zw)
1761 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1763 This opcode is the inverse of :opcode:`DFRACEXP`.
1767 dst.xy = src0.xy \times 2^{src1.xy}
1769 dst.zw = src0.zw \times 2^{src1.zw}
1771 .. opcode:: DMIN - Minimum
1775 dst.xy = min(src0.xy, src1.xy)
1777 dst.zw = min(src0.zw, src1.zw)
1779 .. opcode:: DMAX - Maximum
1783 dst.xy = max(src0.xy, src1.xy)
1785 dst.zw = max(src0.zw, src1.zw)
1787 .. opcode:: DMUL - Multiply
1791 dst.xy = src0.xy \times src1.xy
1793 dst.zw = src0.zw \times src1.zw
1796 .. opcode:: DMAD - Multiply And Add
1800 dst.xy = src0.xy \times src1.xy + src2.xy
1802 dst.zw = src0.zw \times src1.zw + src2.zw
1805 .. opcode:: DRCP - Reciprocal
1809 dst.xy = \frac{1}{src.xy}
1811 dst.zw = \frac{1}{src.zw}
1813 .. opcode:: DSQRT - Square Root
1817 dst.xy = \sqrt{src.xy}
1819 dst.zw = \sqrt{src.zw}
1822 .. _samplingopcodes:
1824 Resource Sampling Opcodes
1825 ^^^^^^^^^^^^^^^^^^^^^^^^^
1827 Those opcodes follow very closely semantics of the respective Direct3D
1828 instructions. If in doubt double check Direct3D documentation.
1829 Note that the swizzle on SVIEW (src1) determines texel swizzling
1832 .. opcode:: SAMPLE - Using provided address, sample data from the
1833 specified texture using the filtering mode identified
1834 by the gven sampler. The source data may come from
1835 any resource type other than buffers.
1836 SAMPLE dst, address, sampler_view, sampler
1838 SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]
1840 .. opcode:: SAMPLE_I - Simplified alternative to the SAMPLE instruction.
1841 Using the provided integer address, SAMPLE_I fetches data
1842 from the specified sampler view without any filtering.
1843 The source data may come from any resource type other
1845 SAMPLE_I dst, address, sampler_view
1847 SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]
1848 The 'address' is specified as unsigned integers. If the
1849 'address' is out of range [0...(# texels - 1)] the
1850 result of the fetch is always 0 in all components.
1851 As such the instruction doesn't honor address wrap
1852 modes, in cases where that behavior is desirable
1853 'SAMPLE' instruction should be used.
1854 address.w always provides an unsigned integer mipmap
1855 level. If the value is out of the range then the
1856 instruction always returns 0 in all components.
1857 address.yz are ignored for buffers and 1d textures.
1858 address.z is ignored for 1d texture arrays and 2d
1860 For 1D texture arrays address.y provides the array
1861 index (also as unsigned integer). If the value is
1862 out of the range of available array indices
1863 [0... (array size - 1)] then the opcode always returns
1864 0 in all components.
1865 For 2D texture arrays address.z provides the array
1866 index, otherwise it exhibits the same behavior as in
1867 the case for 1D texture arrays.
1868 The exact semantics of the source address are presented
1870 resource type X Y Z W
1871 ------------- ------------------------
1872 PIPE_BUFFER x ignored
1873 PIPE_TEXTURE_1D x mpl
1874 PIPE_TEXTURE_2D x y mpl
1875 PIPE_TEXTURE_3D x y z mpl
1876 PIPE_TEXTURE_RECT x y mpl
1877 PIPE_TEXTURE_CUBE not allowed as source
1878 PIPE_TEXTURE_1D_ARRAY x idx mpl
1879 PIPE_TEXTURE_2D_ARRAY x y idx mpl
1881 Where 'mpl' is a mipmap level and 'idx' is the
1884 .. opcode:: SAMPLE_I_MS - Just like SAMPLE_I but allows fetch data from
1885 multi-sampled surfaces.
1886 SAMPLE_I_MS dst, address, sampler_view, sample
1888 .. opcode:: SAMPLE_B - Just like the SAMPLE instruction with the
1889 exception that an additional bias is applied to the
1890 level of detail computed as part of the instruction
1892 SAMPLE_B dst, address, sampler_view, sampler, lod_bias
1894 SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x
1896 .. opcode:: SAMPLE_C - Similar to the SAMPLE instruction but it
1897 performs a comparison filter. The operands to SAMPLE_C
1898 are identical to SAMPLE, except that there is an additional
1899 float32 operand, reference value, which must be a register
1900 with single-component, or a scalar literal.
1901 SAMPLE_C makes the hardware use the current samplers
1902 compare_func (in pipe_sampler_state) to compare
1903 reference value against the red component value for the
1904 surce resource at each texel that the currently configured
1905 texture filter covers based on the provided coordinates.
1906 SAMPLE_C dst, address, sampler_view.r, sampler, ref_value
1908 SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x
1910 .. opcode:: SAMPLE_C_LZ - Same as SAMPLE_C, but LOD is 0 and derivatives
1911 are ignored. The LZ stands for level-zero.
1912 SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value
1914 SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x
1917 .. opcode:: SAMPLE_D - SAMPLE_D is identical to the SAMPLE opcode except
1918 that the derivatives for the source address in the x
1919 direction and the y direction are provided by extra
1921 SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y
1923 SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]
1925 .. opcode:: SAMPLE_L - SAMPLE_L is identical to the SAMPLE opcode except
1926 that the LOD is provided directly as a scalar value,
1927 representing no anisotropy.
1928 SAMPLE_L dst, address, sampler_view, sampler, explicit_lod
1930 SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x
1932 .. opcode:: GATHER4 - Gathers the four texels to be used in a bi-linear
1933 filtering operation and packs them into a single register.
1934 Only works with 2D, 2D array, cubemaps, and cubemaps arrays.
1935 For 2D textures, only the addressing modes of the sampler and
1936 the top level of any mip pyramid are used. Set W to zero.
1937 It behaves like the SAMPLE instruction, but a filtered
1938 sample is not generated. The four samples that contribute
1939 to filtering are placed into xyzw in counter-clockwise order,
1940 starting with the (u,v) texture coordinate delta at the
1941 following locations (-, +), (+, +), (+, -), (-, -), where
1942 the magnitude of the deltas are half a texel.
1945 .. opcode:: SVIEWINFO - query the dimensions of a given sampler view.
1946 dst receives width, height, depth or array size and
1947 number of mipmap levels as int4. The dst can have a writemask
1948 which will specify what info is the caller interested
1950 SVIEWINFO dst, src_mip_level, sampler_view
1952 SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]
1953 src_mip_level is an unsigned integer scalar. If it's
1954 out of range then returns 0 for width, height and
1955 depth/array size but the total number of mipmap is
1956 still returned correctly for the given sampler view.
1957 The returned width, height and depth values are for
1958 the mipmap level selected by the src_mip_level and
1959 are in the number of texels.
1960 For 1d texture array width is in dst.x, array size
1961 is in dst.y and dst.z is 0. The number of mipmaps
1963 In contrast to d3d10 resinfo, there's no way in the
1964 tgsi instruction encoding to specify the return type
1965 (float/rcpfloat/uint), hence always using uint. Also,
1966 unlike the SAMPLE instructions, the swizzle on src1
1967 resinfo allowing swizzling dst values is ignored (due
1968 to the interaction with rcpfloat modifier which requires
1969 some swizzle handling in the state tracker anyway).
1971 .. opcode:: SAMPLE_POS - query the position of a given sample.
1972 dst receives float4 (x, y, 0, 0) indicated where the
1973 sample is located. If the resource is not a multi-sample
1974 resource and not a render target, the result is 0.
1976 .. opcode:: SAMPLE_INFO - dst receives number of samples in x.
1977 If the resource is not a multi-sample resource and
1978 not a render target, the result is 0.
1981 .. _resourceopcodes:
1983 Resource Access Opcodes
1984 ^^^^^^^^^^^^^^^^^^^^^^^
1986 .. opcode:: LOAD - Fetch data from a shader resource
1988 Syntax: ``LOAD dst, resource, address``
1990 Example: ``LOAD TEMP[0], RES[0], TEMP[1]``
1992 Using the provided integer address, LOAD fetches data
1993 from the specified buffer or texture without any
1996 The 'address' is specified as a vector of unsigned
1997 integers. If the 'address' is out of range the result
2000 Only the first mipmap level of a resource can be read
2001 from using this instruction.
2003 For 1D or 2D texture arrays, the array index is
2004 provided as an unsigned integer in address.y or
2005 address.z, respectively. address.yz are ignored for
2006 buffers and 1D textures. address.z is ignored for 1D
2007 texture arrays and 2D textures. address.w is always
2010 .. opcode:: STORE - Write data to a shader resource
2012 Syntax: ``STORE resource, address, src``
2014 Example: ``STORE RES[0], TEMP[0], TEMP[1]``
2016 Using the provided integer address, STORE writes data
2017 to the specified buffer or texture.
2019 The 'address' is specified as a vector of unsigned
2020 integers. If the 'address' is out of range the result
2023 Only the first mipmap level of a resource can be
2024 written to using this instruction.
2026 For 1D or 2D texture arrays, the array index is
2027 provided as an unsigned integer in address.y or
2028 address.z, respectively. address.yz are ignored for
2029 buffers and 1D textures. address.z is ignored for 1D
2030 texture arrays and 2D textures. address.w is always
2034 .. _threadsyncopcodes:
2036 Inter-thread synchronization opcodes
2037 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2039 These opcodes are intended for communication between threads running
2040 within the same compute grid. For now they're only valid in compute
2043 .. opcode:: MFENCE - Memory fence
2045 Syntax: ``MFENCE resource``
2047 Example: ``MFENCE RES[0]``
2049 This opcode forces strong ordering between any memory access
2050 operations that affect the specified resource. This means that
2051 previous loads and stores (and only those) will be performed and
2052 visible to other threads before the program execution continues.
2055 .. opcode:: LFENCE - Load memory fence
2057 Syntax: ``LFENCE resource``
2059 Example: ``LFENCE RES[0]``
2061 Similar to MFENCE, but it only affects the ordering of memory loads.
2064 .. opcode:: SFENCE - Store memory fence
2066 Syntax: ``SFENCE resource``
2068 Example: ``SFENCE RES[0]``
2070 Similar to MFENCE, but it only affects the ordering of memory stores.
2073 .. opcode:: BARRIER - Thread group barrier
2077 This opcode suspends the execution of the current thread until all
2078 the remaining threads in the working group reach the same point of
2079 the program. Results are unspecified if any of the remaining
2080 threads terminates or never reaches an executed BARRIER instruction.
2088 These opcodes provide atomic variants of some common arithmetic and
2089 logical operations. In this context atomicity means that another
2090 concurrent memory access operation that affects the same memory
2091 location is guaranteed to be performed strictly before or after the
2092 entire execution of the atomic operation.
2094 For the moment they're only valid in compute programs.
2096 .. opcode:: ATOMUADD - Atomic integer addition
2098 Syntax: ``ATOMUADD dst, resource, offset, src``
2100 Example: ``ATOMUADD TEMP[0], RES[0], TEMP[1], TEMP[2]``
2102 The following operation is performed atomically on each component:
2106 dst_i = resource[offset]_i
2108 resource[offset]_i = dst_i + src_i
2111 .. opcode:: ATOMXCHG - Atomic exchange
2113 Syntax: ``ATOMXCHG dst, resource, offset, src``
2115 Example: ``ATOMXCHG TEMP[0], RES[0], TEMP[1], TEMP[2]``
2117 The following operation is performed atomically on each component:
2121 dst_i = resource[offset]_i
2123 resource[offset]_i = src_i
2126 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2128 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2130 Example: ``ATOMCAS TEMP[0], RES[0], TEMP[1], TEMP[2], TEMP[3]``
2132 The following operation is performed atomically on each component:
2136 dst_i = resource[offset]_i
2138 resource[offset]_i = (dst_i == cmp_i ? src_i : dst_i)
2141 .. opcode:: ATOMAND - Atomic bitwise And
2143 Syntax: ``ATOMAND dst, resource, offset, src``
2145 Example: ``ATOMAND 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:: ATOMOR - Atomic bitwise Or
2158 Syntax: ``ATOMOR dst, resource, offset, src``
2160 Example: ``ATOMOR 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 = dst_i | src_i
2171 .. opcode:: ATOMXOR - Atomic bitwise Xor
2173 Syntax: ``ATOMXOR dst, resource, offset, src``
2175 Example: ``ATOMXOR TEMP[0], RES[0], TEMP[1], TEMP[2]``
2177 The following operation is performed atomically on each component:
2181 dst_i = resource[offset]_i
2183 resource[offset]_i = dst_i \oplus src_i
2186 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2188 Syntax: ``ATOMUMIN dst, resource, offset, src``
2190 Example: ``ATOMUMIN 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 ? dst_i : src_i)
2201 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2203 Syntax: ``ATOMUMAX dst, resource, offset, src``
2205 Example: ``ATOMUMAX 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 ? dst_i : src_i)
2216 .. opcode:: ATOMIMIN - Atomic signed minimum
2218 Syntax: ``ATOMIMIN dst, resource, offset, src``
2220 Example: ``ATOMIMIN 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 < src_i ? dst_i : src_i)
2231 .. opcode:: ATOMIMAX - Atomic signed maximum
2233 Syntax: ``ATOMIMAX dst, resource, offset, src``
2235 Example: ``ATOMIMAX 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)
2247 Explanation of symbols used
2248 ------------------------------
2255 :math:`|x|` Absolute value of `x`.
2257 :math:`\lceil x \rceil` Ceiling of `x`.
2259 clamp(x,y,z) Clamp x between y and z.
2260 (x < y) ? y : (x > z) ? z : x
2262 :math:`\lfloor x\rfloor` Floor of `x`.
2264 :math:`\log_2{x}` Logarithm of `x`, base 2.
2266 max(x,y) Maximum of x and y.
2269 min(x,y) Minimum of x and y.
2272 partialx(x) Derivative of x relative to fragment's X.
2274 partialy(x) Derivative of x relative to fragment's Y.
2276 pop() Pop from stack.
2278 :math:`x^y` `x` to the power `y`.
2280 push(x) Push x on stack.
2284 trunc(x) Truncate x, i.e. drop the fraction bits.
2291 discard Discard fragment.
2295 target Label of target instruction.
2306 Declares a register that is will be referenced as an operand in Instruction
2309 File field contains register file that is being declared and is one
2312 UsageMask field specifies which of the register components can be accessed
2313 and is one of TGSI_WRITEMASK.
2315 The Local flag specifies that a given value isn't intended for
2316 subroutine parameter passing and, as a result, the implementation
2317 isn't required to give any guarantees of it being preserved across
2318 subroutine boundaries. As it's merely a compiler hint, the
2319 implementation is free to ignore it.
2321 If Dimension flag is set to 1, a Declaration Dimension token follows.
2323 If Semantic flag is set to 1, a Declaration Semantic token follows.
2325 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
2327 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
2329 If Array flag is set to 1, a Declaration Array token follows.
2332 ^^^^^^^^^^^^^^^^^^^^^^^^
2334 Declarations can optional have an ArrayID attribute which can be referred by
2335 indirect addressing operands. An ArrayID of zero is reserved and treaded as
2336 if no ArrayID is specified.
2338 If an indirect addressing operand refers to a specific declaration by using
2339 an ArrayID only the registers in this declaration are guaranteed to be
2340 accessed, accessing any register outside this declaration results in undefined
2341 behavior. Note that for compatibility the effective index is zero-based and
2342 not relative to the specified declaration
2344 If no ArrayID is specified with an indirect addressing operand the whole
2345 register file might be accessed by this operand. This is strongly discouraged
2346 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
2348 Declaration Semantic
2349 ^^^^^^^^^^^^^^^^^^^^^^^^
2351 Vertex and fragment shader input and output registers may be labeled
2352 with semantic information consisting of a name and index.
2354 Follows Declaration token if Semantic bit is set.
2356 Since its purpose is to link a shader with other stages of the pipeline,
2357 it is valid to follow only those Declaration tokens that declare a register
2358 either in INPUT or OUTPUT file.
2360 SemanticName field contains the semantic name of the register being declared.
2361 There is no default value.
2363 SemanticIndex is an optional subscript that can be used to distinguish
2364 different register declarations with the same semantic name. The default value
2367 The meanings of the individual semantic names are explained in the following
2370 TGSI_SEMANTIC_POSITION
2371 """"""""""""""""""""""
2373 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
2374 output register which contains the homogeneous vertex position in the clip
2375 space coordinate system. After clipping, the X, Y and Z components of the
2376 vertex will be divided by the W value to get normalized device coordinates.
2378 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
2379 fragment shader input contains the fragment's window position. The X
2380 component starts at zero and always increases from left to right.
2381 The Y component starts at zero and always increases but Y=0 may either
2382 indicate the top of the window or the bottom depending on the fragment
2383 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
2384 The Z coordinate ranges from 0 to 1 to represent depth from the front
2385 to the back of the Z buffer. The W component contains the reciprocol
2386 of the interpolated vertex position W component.
2388 Fragment shaders may also declare an output register with
2389 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
2390 the fragment shader to change the fragment's Z position.
2397 For vertex shader outputs or fragment shader inputs/outputs, this
2398 label indicates that the resister contains an R,G,B,A color.
2400 Several shader inputs/outputs may contain colors so the semantic index
2401 is used to distinguish them. For example, color[0] may be the diffuse
2402 color while color[1] may be the specular color.
2404 This label is needed so that the flat/smooth shading can be applied
2405 to the right interpolants during rasterization.
2409 TGSI_SEMANTIC_BCOLOR
2410 """"""""""""""""""""
2412 Back-facing colors are only used for back-facing polygons, and are only valid
2413 in vertex shader outputs. After rasterization, all polygons are front-facing
2414 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
2415 so all BCOLORs effectively become regular COLORs in the fragment shader.
2421 Vertex shader inputs and outputs and fragment shader inputs may be
2422 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
2423 a fog coordinate. Typically, the fragment shader will use the fog coordinate
2424 to compute a fog blend factor which is used to blend the normal fragment color
2425 with a constant fog color. But fog coord really is just an ordinary vec4
2426 register like regular semantics.
2432 Vertex shader input and output registers may be labeled with
2433 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
2434 in the form (S, 0, 0, 1). The point size controls the width or diameter
2435 of points for rasterization. This label cannot be used in fragment
2438 When using this semantic, be sure to set the appropriate state in the
2439 :ref:`rasterizer` first.
2442 TGSI_SEMANTIC_TEXCOORD
2443 """"""""""""""""""""""
2445 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2447 Vertex shader outputs and fragment shader inputs may be labeled with
2448 this semantic to make them replaceable by sprite coordinates via the
2449 sprite_coord_enable state in the :ref:`rasterizer`.
2450 The semantic index permitted with this semantic is limited to <= 7.
2452 If the driver does not support TEXCOORD, sprite coordinate replacement
2453 applies to inputs with the GENERIC semantic instead.
2455 The intended use case for this semantic is gl_TexCoord.
2458 TGSI_SEMANTIC_PCOORD
2459 """"""""""""""""""""
2461 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
2463 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
2464 that the register contains sprite coordinates in the form (x, y, 0, 1), if
2465 the current primitive is a point and point sprites are enabled. Otherwise,
2466 the contents of the register are undefined.
2468 The intended use case for this semantic is gl_PointCoord.
2471 TGSI_SEMANTIC_GENERIC
2472 """""""""""""""""""""
2474 All vertex/fragment shader inputs/outputs not labeled with any other
2475 semantic label can be considered to be generic attributes. Typical
2476 uses of generic inputs/outputs are texcoords and user-defined values.
2479 TGSI_SEMANTIC_NORMAL
2480 """"""""""""""""""""
2482 Indicates that a vertex shader input is a normal vector. This is
2483 typically only used for legacy graphics APIs.
2489 This label applies to fragment shader inputs only and indicates that
2490 the register contains front/back-face information of the form (F, 0,
2491 0, 1). The first component will be positive when the fragment belongs
2492 to a front-facing polygon, and negative when the fragment belongs to a
2493 back-facing polygon.
2496 TGSI_SEMANTIC_EDGEFLAG
2497 """"""""""""""""""""""
2499 For vertex shaders, this sematic label indicates that an input or
2500 output is a boolean edge flag. The register layout is [F, x, x, x]
2501 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
2502 simply copies the edge flag input to the edgeflag output.
2504 Edge flags are used to control which lines or points are actually
2505 drawn when the polygon mode converts triangles/quads/polygons into
2509 TGSI_SEMANTIC_STENCIL
2510 """""""""""""""""""""
2512 For fragment shaders, this semantic label indicates that an output
2513 is a writable stencil reference value. Only the Y component is writable.
2514 This allows the fragment shader to change the fragments stencilref value.
2517 TGSI_SEMANTIC_VIEWPORT_INDEX
2518 """"""""""""""""""""""""""""
2520 For geometry shaders, this semantic label indicates that an output
2521 contains the index of the viewport (and scissor) to use.
2522 Only the X value is used.
2528 For geometry shaders, this semantic label indicates that an output
2529 contains the layer value to use for the color and depth/stencil surfaces.
2530 Only the X value is used. (Also known as rendertarget array index.)
2533 TGSI_SEMANTIC_CULLDIST
2534 """"""""""""""""""""""
2536 Used as distance to plane for performing application-defined culling
2537 of individual primitives against a plane. When components of vertex
2538 elements are given this label, these values are assumed to be a
2539 float32 signed distance to a plane. Primitives will be completely
2540 discarded if the plane distance for all of the vertices in the
2541 primitive are < 0. If a vertex has a cull distance of NaN, that
2542 vertex counts as "out" (as if its < 0);
2543 The limits on both clip and cull distances are bound
2544 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2545 the maximum number of components that can be used to hold the
2546 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2547 which specifies the maximum number of registers which can be
2548 annotated with those semantics.
2551 TGSI_SEMANTIC_CLIPDIST
2552 """"""""""""""""""""""
2554 When components of vertex elements are identified this way, these
2555 values are each assumed to be a float32 signed distance to a plane.
2556 Primitive setup only invokes rasterization on pixels for which
2557 the interpolated plane distances are >= 0. Multiple clip planes
2558 can be implemented simultaneously, by annotating multiple
2559 components of one or more vertex elements with the above specified
2560 semantic. The limits on both clip and cull distances are bound
2561 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
2562 the maximum number of components that can be used to hold the
2563 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
2564 which specifies the maximum number of registers which can be
2565 annotated with those semantics.
2568 Declaration Interpolate
2569 ^^^^^^^^^^^^^^^^^^^^^^^
2571 This token is only valid for fragment shader INPUT declarations.
2573 The Interpolate field specifes the way input is being interpolated by
2574 the rasteriser and is one of TGSI_INTERPOLATE_*.
2576 The CylindricalWrap bitfield specifies which register components
2577 should be subject to cylindrical wrapping when interpolating by the
2578 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
2579 should be interpolated according to cylindrical wrapping rules.
2582 Declaration Sampler View
2583 ^^^^^^^^^^^^^^^^^^^^^^^^
2585 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
2587 DCL SVIEW[#], resource, type(s)
2589 Declares a shader input sampler view and assigns it to a SVIEW[#]
2592 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
2594 type must be 1 or 4 entries (if specifying on a per-component
2595 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
2598 Declaration Resource
2599 ^^^^^^^^^^^^^^^^^^^^
2601 Follows Declaration token if file is TGSI_FILE_RESOURCE.
2603 DCL RES[#], resource [, WR] [, RAW]
2605 Declares a shader input resource and assigns it to a RES[#]
2608 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
2611 If the RAW keyword is not specified, the texture data will be
2612 subject to conversion, swizzling and scaling as required to yield
2613 the specified data type from the physical data format of the bound
2616 If the RAW keyword is specified, no channel conversion will be
2617 performed: the values read for each of the channels (X,Y,Z,W) will
2618 correspond to consecutive words in the same order and format
2619 they're found in memory. No element-to-address conversion will be
2620 performed either: the value of the provided X coordinate will be
2621 interpreted in byte units instead of texel units. The result of
2622 accessing a misaligned address is undefined.
2624 Usage of the STORE opcode is only allowed if the WR (writable) flag
2629 ^^^^^^^^^^^^^^^^^^^^^^^^
2632 Properties are general directives that apply to the whole TGSI program.
2637 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
2638 The default value is UPPER_LEFT.
2640 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
2641 increase downward and rightward.
2642 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
2643 increase upward and rightward.
2645 OpenGL defaults to LOWER_LEFT, and is configurable with the
2646 GL_ARB_fragment_coord_conventions extension.
2648 DirectX 9/10 use UPPER_LEFT.
2650 FS_COORD_PIXEL_CENTER
2651 """""""""""""""""""""
2653 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
2654 The default value is HALF_INTEGER.
2656 If HALF_INTEGER, the fractionary part of the position will be 0.5
2657 If INTEGER, the fractionary part of the position will be 0.0
2659 Note that this does not affect the set of fragments generated by
2660 rasterization, which is instead controlled by half_pixel_center in the
2663 OpenGL defaults to HALF_INTEGER, and is configurable with the
2664 GL_ARB_fragment_coord_conventions extension.
2666 DirectX 9 uses INTEGER.
2667 DirectX 10 uses HALF_INTEGER.
2669 FS_COLOR0_WRITES_ALL_CBUFS
2670 """"""""""""""""""""""""""
2671 Specifies that writes to the fragment shader color 0 are replicated to all
2672 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
2673 fragData is directed to a single color buffer, but fragColor is broadcast.
2676 """"""""""""""""""""""""""
2677 If this property is set on the program bound to the shader stage before the
2678 fragment shader, user clip planes should have no effect (be disabled) even if
2679 that shader does not write to any clip distance outputs and the rasterizer's
2680 clip_plane_enable is non-zero.
2681 This property is only supported by drivers that also support shader clip
2683 This is useful for APIs that don't have UCPs and where clip distances written
2684 by a shader cannot be disabled.
2687 Texture Sampling and Texture Formats
2688 ------------------------------------
2690 This table shows how texture image components are returned as (x,y,z,w) tuples
2691 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
2692 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
2695 +--------------------+--------------+--------------------+--------------+
2696 | Texture Components | Gallium | OpenGL | Direct3D 9 |
2697 +====================+==============+====================+==============+
2698 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
2699 +--------------------+--------------+--------------------+--------------+
2700 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
2701 +--------------------+--------------+--------------------+--------------+
2702 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
2703 +--------------------+--------------+--------------------+--------------+
2704 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
2705 +--------------------+--------------+--------------------+--------------+
2706 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
2707 +--------------------+--------------+--------------------+--------------+
2708 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
2709 +--------------------+--------------+--------------------+--------------+
2710 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
2711 +--------------------+--------------+--------------------+--------------+
2712 | I | (i, i, i, i) | (i, i, i, i) | N/A |
2713 +--------------------+--------------+--------------------+--------------+
2714 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
2715 | | | [#envmap-bumpmap]_ | |
2716 +--------------------+--------------+--------------------+--------------+
2717 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
2718 | | | [#depth-tex-mode]_ | |
2719 +--------------------+--------------+--------------------+--------------+
2720 | S | (s, s, s, s) | unknown | unknown |
2721 +--------------------+--------------+--------------------+--------------+
2723 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
2724 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
2725 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.