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 and precise modifier
32 For arithmetic instruction having a precise modifier certain optimizations
33 which may alter the result are disallowed. Example: *add(mul(a,b),c)* can't be
34 optimized to TGSI_OPCODE_MAD, because some hardware only supports the fused
37 For inputs which have a floating point type, both absolute value and
38 negation modifiers are supported (with absolute value being applied
39 first). The only source of TGSI_OPCODE_MOV and the second and third
40 sources of TGSI_OPCODE_UCMP are considered to have float type for
43 For inputs which have signed or unsigned type only the negate modifier is
50 ^^^^^^^^^^^^^^^^^^^^^^^^^
52 These opcodes are guaranteed to be available regardless of the driver being
55 .. opcode:: ARL - Address Register Load
59 dst.x = (int) \lfloor src.x\rfloor
61 dst.y = (int) \lfloor src.y\rfloor
63 dst.z = (int) \lfloor src.z\rfloor
65 dst.w = (int) \lfloor src.w\rfloor
68 .. opcode:: MOV - Move
81 .. opcode:: LIT - Light Coefficients
86 dst.y &= max(src.x, 0) \\
87 dst.z &= (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0 \\
91 .. opcode:: RCP - Reciprocal
93 This instruction replicates its result.
100 .. opcode:: RSQ - Reciprocal Square Root
102 This instruction replicates its result. The results are undefined for src <= 0.
106 dst = \frac{1}{\sqrt{src.x}}
109 .. opcode:: SQRT - Square Root
111 This instruction replicates its result. The results are undefined for src < 0.
118 .. opcode:: EXP - Approximate Exponential Base 2
122 dst.x &= 2^{\lfloor src.x\rfloor} \\
123 dst.y &= src.x - \lfloor src.x\rfloor \\
124 dst.z &= 2^{src.x} \\
128 .. opcode:: LOG - Approximate Logarithm Base 2
132 dst.x &= \lfloor\log_2{|src.x|}\rfloor \\
133 dst.y &= \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}} \\
134 dst.z &= \log_2{|src.x|} \\
138 .. opcode:: MUL - Multiply
142 dst.x = src0.x \times src1.x
144 dst.y = src0.y \times src1.y
146 dst.z = src0.z \times src1.z
148 dst.w = src0.w \times src1.w
151 .. opcode:: ADD - Add
155 dst.x = src0.x + src1.x
157 dst.y = src0.y + src1.y
159 dst.z = src0.z + src1.z
161 dst.w = src0.w + src1.w
164 .. opcode:: DP3 - 3-component Dot Product
166 This instruction replicates its result.
170 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z
173 .. opcode:: DP4 - 4-component Dot Product
175 This instruction replicates its result.
179 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src0.w \times src1.w
182 .. opcode:: DST - Distance Vector
187 dst.y &= src0.y \times src1.y\\
192 .. opcode:: MIN - Minimum
196 dst.x = min(src0.x, src1.x)
198 dst.y = min(src0.y, src1.y)
200 dst.z = min(src0.z, src1.z)
202 dst.w = min(src0.w, src1.w)
205 .. opcode:: MAX - Maximum
209 dst.x = max(src0.x, src1.x)
211 dst.y = max(src0.y, src1.y)
213 dst.z = max(src0.z, src1.z)
215 dst.w = max(src0.w, src1.w)
218 .. opcode:: SLT - Set On Less Than
222 dst.x = (src0.x < src1.x) ? 1.0F : 0.0F
224 dst.y = (src0.y < src1.y) ? 1.0F : 0.0F
226 dst.z = (src0.z < src1.z) ? 1.0F : 0.0F
228 dst.w = (src0.w < src1.w) ? 1.0F : 0.0F
231 .. opcode:: SGE - Set On Greater Equal Than
235 dst.x = (src0.x >= src1.x) ? 1.0F : 0.0F
237 dst.y = (src0.y >= src1.y) ? 1.0F : 0.0F
239 dst.z = (src0.z >= src1.z) ? 1.0F : 0.0F
241 dst.w = (src0.w >= src1.w) ? 1.0F : 0.0F
244 .. opcode:: MAD - Multiply And Add
246 Perform a * b + c. The implementation is free to decide whether there is an
247 intermediate rounding step or not.
251 dst.x = src0.x \times src1.x + src2.x
253 dst.y = src0.y \times src1.y + src2.y
255 dst.z = src0.z \times src1.z + src2.z
257 dst.w = src0.w \times src1.w + src2.w
260 .. opcode:: LRP - Linear Interpolate
264 dst.x = src0.x \times src1.x + (1 - src0.x) \times src2.x
266 dst.y = src0.y \times src1.y + (1 - src0.y) \times src2.y
268 dst.z = src0.z \times src1.z + (1 - src0.z) \times src2.z
270 dst.w = src0.w \times src1.w + (1 - src0.w) \times src2.w
273 .. opcode:: FMA - Fused Multiply-Add
275 Perform a * b + c with no intermediate rounding step.
279 dst.x = src0.x \times src1.x + src2.x
281 dst.y = src0.y \times src1.y + src2.y
283 dst.z = src0.z \times src1.z + src2.z
285 dst.w = src0.w \times src1.w + src2.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:: FLR - Floor
318 dst.x = \lfloor src.x\rfloor
320 dst.y = \lfloor src.y\rfloor
322 dst.z = \lfloor src.z\rfloor
324 dst.w = \lfloor src.w\rfloor
327 .. opcode:: ROUND - Round
340 .. opcode:: EX2 - Exponential Base 2
342 This instruction replicates its result.
349 .. opcode:: LG2 - Logarithm Base 2
351 This instruction replicates its result.
358 .. opcode:: POW - Power
360 This instruction replicates its result.
364 dst = src0.x^{src1.x}
366 .. opcode:: XPD - Cross Product
370 dst.x = src0.y \times src1.z - src1.y \times src0.z
372 dst.y = src0.z \times src1.x - src1.z \times src0.x
374 dst.z = src0.x \times src1.y - src1.x \times src0.y
379 .. opcode:: DPH - Homogeneous Dot Product
381 This instruction replicates its result.
385 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
388 .. opcode:: COS - Cosine
390 This instruction replicates its result.
397 .. opcode:: DDX, DDX_FINE - Derivative Relative To X
399 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
400 advertised. When it is, the fine version guarantees one derivative per row
401 while DDX is allowed to be the same for the entire 2x2 quad.
405 dst.x = partialx(src.x)
407 dst.y = partialx(src.y)
409 dst.z = partialx(src.z)
411 dst.w = partialx(src.w)
414 .. opcode:: DDY, DDY_FINE - Derivative Relative To Y
416 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
417 advertised. When it is, the fine version guarantees one derivative per column
418 while DDY is allowed to be the same for the entire 2x2 quad.
422 dst.x = partialy(src.x)
424 dst.y = partialy(src.y)
426 dst.z = partialy(src.z)
428 dst.w = partialy(src.w)
431 .. opcode:: PK2H - Pack Two 16-bit Floats
433 This instruction replicates its result.
437 dst = f32\_to\_f16(src.x) | f32\_to\_f16(src.y) << 16
440 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
445 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
450 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
455 .. opcode:: SEQ - Set On Equal
459 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
461 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
463 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
465 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
468 .. opcode:: SGT - Set On Greater Than
472 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
474 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
476 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
478 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
481 .. opcode:: SIN - Sine
483 This instruction replicates its result.
490 .. opcode:: SLE - Set On Less Equal Than
494 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
496 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
498 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
500 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
503 .. opcode:: SNE - Set On Not Equal
507 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
509 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
511 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
513 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
516 .. opcode:: TEX - Texture Lookup
518 for array textures src0.y contains the slice for 1D,
519 and src0.z contain the slice for 2D.
521 for shadow textures with no arrays (and not cube map),
522 src0.z contains the reference value.
524 for shadow textures with arrays, src0.z contains
525 the reference value for 1D arrays, and src0.w contains
526 the reference value for 2D arrays and cube maps.
528 for cube map array shadow textures, the reference value
529 cannot be passed in src0.w, and TEX2 must be used instead.
535 shadow_ref = src0.z or src0.w (optional)
539 dst = texture\_sample(unit, coord, shadow_ref)
542 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
544 this is the same as TEX, but uses another reg to encode the
555 dst = texture\_sample(unit, coord, shadow_ref)
560 .. opcode:: TXD - Texture Lookup with Derivatives
572 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
575 .. opcode:: TXP - Projective Texture Lookup
579 coord.x = src0.x / src0.w
581 coord.y = src0.y / src0.w
583 coord.z = src0.z / src0.w
589 dst = texture\_sample(unit, coord)
592 .. opcode:: UP2H - Unpack Two 16-Bit Floats
596 dst.x = f16\_to\_f32(src0.x \& 0xffff)
598 dst.y = f16\_to\_f32(src0.x >> 16)
600 dst.z = f16\_to\_f32(src0.x \& 0xffff)
602 dst.w = f16\_to\_f32(src0.x >> 16)
606 Considered for removal.
608 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
614 Considered for removal.
616 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
622 Considered for removal.
624 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
630 Considered for removal.
633 .. opcode:: ARR - Address Register Load With Round
637 dst.x = (int) round(src.x)
639 dst.y = (int) round(src.y)
641 dst.z = (int) round(src.z)
643 dst.w = (int) round(src.w)
646 .. opcode:: SSG - Set Sign
650 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
652 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
654 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
656 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
659 .. opcode:: CMP - Compare
663 dst.x = (src0.x < 0) ? src1.x : src2.x
665 dst.y = (src0.y < 0) ? src1.y : src2.y
667 dst.z = (src0.z < 0) ? src1.z : src2.z
669 dst.w = (src0.w < 0) ? src1.w : src2.w
672 .. opcode:: KILL_IF - Conditional Discard
674 Conditional discard. Allowed in fragment shaders only.
678 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
683 .. opcode:: KILL - Discard
685 Unconditional discard. Allowed in fragment shaders only.
688 .. opcode:: SCS - Sine Cosine
701 .. opcode:: TXB - Texture Lookup With Bias
703 for cube map array textures and shadow cube maps, the bias value
704 cannot be passed in src0.w, and TXB2 must be used instead.
706 if the target is a shadow texture, the reference value is always
707 in src.z (this prevents shadow 3d and shadow 2d arrays from
708 using this instruction, but this is not needed).
724 dst = texture\_sample(unit, coord, bias)
727 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
729 this is the same as TXB, but uses another reg to encode the
730 lod bias value for cube map arrays and shadow cube maps.
731 Presumably shadow 2d arrays and shadow 3d targets could use
732 this encoding too, but this is not legal.
734 shadow cube map arrays are neither possible nor required.
744 dst = texture\_sample(unit, coord, bias)
747 .. opcode:: DIV - Divide
751 dst.x = \frac{src0.x}{src1.x}
753 dst.y = \frac{src0.y}{src1.y}
755 dst.z = \frac{src0.z}{src1.z}
757 dst.w = \frac{src0.w}{src1.w}
760 .. opcode:: DP2 - 2-component Dot Product
762 This instruction replicates its result.
766 dst = src0.x \times src1.x + src0.y \times src1.y
769 .. opcode:: TEX_LZ - Texture Lookup With LOD = 0
771 This is the same as TXL with LOD = 0. Like every texture opcode, it obeys
772 pipe_sampler_view::u.tex.first_level and pipe_sampler_state::min_lod.
773 There is no way to override those two in shaders.
789 dst = texture\_sample(unit, coord, lod)
792 .. opcode:: TXL - Texture Lookup With explicit LOD
794 for cube map array textures, the explicit lod value
795 cannot be passed in src0.w, and TXL2 must be used instead.
797 if the target is a shadow texture, the reference value is always
798 in src.z (this prevents shadow 3d / 2d array / cube targets from
799 using this instruction, but this is not needed).
815 dst = texture\_sample(unit, coord, lod)
818 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
820 this is the same as TXL, but uses another reg to encode the
822 Presumably shadow 3d / 2d array / cube targets could use
823 this encoding too, but this is not legal.
825 shadow cube map arrays are neither possible nor required.
835 dst = texture\_sample(unit, coord, lod)
838 .. opcode:: PUSHA - Push Address Register On Stack
847 Considered for cleanup.
851 Considered for removal.
853 .. opcode:: POPA - Pop Address Register From Stack
862 Considered for cleanup.
866 Considered for removal.
869 .. opcode:: CALLNZ - Subroutine Call If Not Zero
875 Considered for cleanup.
879 Considered for removal.
883 ^^^^^^^^^^^^^^^^^^^^^^^^
885 These opcodes are primarily provided for special-use computational shaders.
886 Support for these opcodes indicated by a special pipe capability bit (TBD).
888 XXX doesn't look like most of the opcodes really belong here.
890 .. opcode:: CEIL - Ceiling
894 dst.x = \lceil src.x\rceil
896 dst.y = \lceil src.y\rceil
898 dst.z = \lceil src.z\rceil
900 dst.w = \lceil src.w\rceil
903 .. opcode:: TRUNC - Truncate
916 .. opcode:: MOD - Modulus
920 dst.x = src0.x \bmod src1.x
922 dst.y = src0.y \bmod src1.y
924 dst.z = src0.z \bmod src1.z
926 dst.w = src0.w \bmod src1.w
929 .. opcode:: UARL - Integer Address Register Load
931 Moves the contents of the source register, assumed to be an integer, into the
932 destination register, which is assumed to be an address (ADDR) register.
935 .. opcode:: SAD - Sum Of Absolute Differences
939 dst.x = |src0.x - src1.x| + src2.x
941 dst.y = |src0.y - src1.y| + src2.y
943 dst.z = |src0.z - src1.z| + src2.z
945 dst.w = |src0.w - src1.w| + src2.w
948 .. opcode:: TXF - Texel Fetch
950 As per NV_gpu_shader4, extract a single texel from a specified texture
951 image or PIPE_BUFFER resource. The source sampler may not be a CUBE or
953 four-component signed integer vector used to identify the single texel
954 accessed. 3 components + level. Just like texture instructions, an optional
955 offset vector is provided, which is subject to various driver restrictions
956 (regarding range, source of offsets). This instruction ignores the sampler
959 TXF(uint_vec coord, int_vec offset).
962 .. opcode:: TXF_LZ - Texel Fetch
964 This is the same as TXF with level = 0. Like TXF, it obeys
965 pipe_sampler_view::u.tex.first_level.
968 .. opcode:: TXQ - Texture Size Query
970 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
971 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
972 depth), 1D array (width, layers), 2D array (width, height, layers).
973 Also return the number of accessible levels (last_level - first_level + 1)
976 For components which don't return a resource dimension, their value
983 dst.x = texture\_width(unit, lod)
985 dst.y = texture\_height(unit, lod)
987 dst.z = texture\_depth(unit, lod)
989 dst.w = texture\_levels(unit)
992 .. opcode:: TXQS - Texture Samples Query
994 This retrieves the number of samples in the texture, and stores it
995 into the x component as an unsigned integer. The other components are
996 undefined. If the texture is not multisampled, this function returns
997 (1, undef, undef, undef).
1001 dst.x = texture\_samples(unit)
1004 .. opcode:: TG4 - Texture Gather
1006 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
1007 filtering operation and packs them into a single register. Only works with
1008 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
1009 addressing modes of the sampler and the top level of any mip pyramid are
1010 used. Set W to zero. It behaves like the TEX instruction, but a filtered
1011 sample is not generated. The four samples that contribute to filtering are
1012 placed into xyzw in clockwise order, starting with the (u,v) texture
1013 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
1014 where the magnitude of the deltas are half a texel.
1016 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
1017 depth compares, single component selection, and a non-constant offset. It
1018 doesn't allow support for the GL independent offset to get i0,j0. This would
1019 require another CAP is hw can do it natively. For now we lower that before
1028 dst = texture\_gather4 (unit, coord, component)
1030 (with SM5 - cube array shadow)
1038 dst = texture\_gather (uint, coord, compare)
1040 .. opcode:: LODQ - level of detail query
1042 Compute the LOD information that the texture pipe would use to access the
1043 texture. The Y component contains the computed LOD lambda_prime. The X
1044 component contains the LOD that will be accessed, based on min/max lod's
1051 dst.xy = lodq(uint, coord);
1053 .. opcode:: CLOCK - retrieve the current shader time
1055 Invoking this instruction multiple times in the same shader should
1056 cause monotonically increasing values to be returned. The values
1057 are implicitly 64-bit, so if fewer than 64 bits of precision are
1058 available, to provide expected wraparound semantics, the value
1059 should be shifted up so that the most significant bit of the time
1060 is the most significant bit of the 64-bit value.
1068 ^^^^^^^^^^^^^^^^^^^^^^^^
1069 These opcodes are used for integer operations.
1070 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1073 .. opcode:: I2F - Signed Integer To Float
1075 Rounding is unspecified (round to nearest even suggested).
1079 dst.x = (float) src.x
1081 dst.y = (float) src.y
1083 dst.z = (float) src.z
1085 dst.w = (float) src.w
1088 .. opcode:: U2F - Unsigned Integer To Float
1090 Rounding is unspecified (round to nearest even suggested).
1094 dst.x = (float) src.x
1096 dst.y = (float) src.y
1098 dst.z = (float) src.z
1100 dst.w = (float) src.w
1103 .. opcode:: F2I - Float to Signed Integer
1105 Rounding is towards zero (truncate).
1106 Values outside signed range (including NaNs) produce undefined results.
1119 .. opcode:: F2U - Float to Unsigned Integer
1121 Rounding is towards zero (truncate).
1122 Values outside unsigned range (including NaNs) produce undefined results.
1126 dst.x = (unsigned) src.x
1128 dst.y = (unsigned) src.y
1130 dst.z = (unsigned) src.z
1132 dst.w = (unsigned) src.w
1135 .. opcode:: UADD - Integer Add
1137 This instruction works the same for signed and unsigned integers.
1138 The low 32bit of the result is returned.
1142 dst.x = src0.x + 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:: UMAD - Integer Multiply And Add
1153 This instruction works the same for signed and unsigned integers.
1154 The multiplication returns the low 32bit (as does the result itself).
1158 dst.x = src0.x \times src1.x + src2.x
1160 dst.y = src0.y \times src1.y + src2.y
1162 dst.z = src0.z \times src1.z + src2.z
1164 dst.w = src0.w \times src1.w + src2.w
1167 .. opcode:: UMUL - Integer Multiply
1169 This instruction works the same for signed and unsigned integers.
1170 The low 32bit of the result is returned.
1174 dst.x = src0.x \times src1.x
1176 dst.y = src0.y \times src1.y
1178 dst.z = src0.z \times src1.z
1180 dst.w = src0.w \times src1.w
1183 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1185 The high 32bits of the multiplication of 2 signed integers are returned.
1189 dst.x = (src0.x \times src1.x) >> 32
1191 dst.y = (src0.y \times src1.y) >> 32
1193 dst.z = (src0.z \times src1.z) >> 32
1195 dst.w = (src0.w \times src1.w) >> 32
1198 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1200 The high 32bits of the multiplication of 2 unsigned integers are returned.
1204 dst.x = (src0.x \times src1.x) >> 32
1206 dst.y = (src0.y \times src1.y) >> 32
1208 dst.z = (src0.z \times src1.z) >> 32
1210 dst.w = (src0.w \times src1.w) >> 32
1213 .. opcode:: IDIV - Signed Integer Division
1215 TBD: behavior for division by zero.
1219 dst.x = \frac{src0.x}{src1.x}
1221 dst.y = \frac{src0.y}{src1.y}
1223 dst.z = \frac{src0.z}{src1.z}
1225 dst.w = \frac{src0.w}{src1.w}
1228 .. opcode:: UDIV - Unsigned Integer Division
1230 For division by zero, 0xffffffff is returned.
1234 dst.x = \frac{src0.x}{src1.x}
1236 dst.y = \frac{src0.y}{src1.y}
1238 dst.z = \frac{src0.z}{src1.z}
1240 dst.w = \frac{src0.w}{src1.w}
1243 .. opcode:: UMOD - Unsigned Integer Remainder
1245 If second arg is zero, 0xffffffff is returned.
1249 dst.x = src0.x \bmod src1.x
1251 dst.y = src0.y \bmod src1.y
1253 dst.z = src0.z \bmod src1.z
1255 dst.w = src0.w \bmod src1.w
1258 .. opcode:: NOT - Bitwise Not
1271 .. opcode:: AND - Bitwise And
1275 dst.x = src0.x \& src1.x
1277 dst.y = src0.y \& src1.y
1279 dst.z = src0.z \& src1.z
1281 dst.w = src0.w \& src1.w
1284 .. opcode:: OR - Bitwise Or
1288 dst.x = src0.x | src1.x
1290 dst.y = src0.y | src1.y
1292 dst.z = src0.z | src1.z
1294 dst.w = src0.w | src1.w
1297 .. opcode:: XOR - Bitwise Xor
1301 dst.x = src0.x \oplus src1.x
1303 dst.y = src0.y \oplus src1.y
1305 dst.z = src0.z \oplus src1.z
1307 dst.w = src0.w \oplus src1.w
1310 .. opcode:: IMAX - Maximum of Signed Integers
1314 dst.x = max(src0.x, src1.x)
1316 dst.y = max(src0.y, src1.y)
1318 dst.z = max(src0.z, src1.z)
1320 dst.w = max(src0.w, src1.w)
1323 .. opcode:: UMAX - Maximum of Unsigned Integers
1327 dst.x = max(src0.x, src1.x)
1329 dst.y = max(src0.y, src1.y)
1331 dst.z = max(src0.z, src1.z)
1333 dst.w = max(src0.w, src1.w)
1336 .. opcode:: IMIN - Minimum of Signed Integers
1340 dst.x = min(src0.x, src1.x)
1342 dst.y = min(src0.y, src1.y)
1344 dst.z = min(src0.z, src1.z)
1346 dst.w = min(src0.w, src1.w)
1349 .. opcode:: UMIN - Minimum of Unsigned Integers
1353 dst.x = min(src0.x, src1.x)
1355 dst.y = min(src0.y, src1.y)
1357 dst.z = min(src0.z, src1.z)
1359 dst.w = min(src0.w, src1.w)
1362 .. opcode:: SHL - Shift Left
1364 The shift count is masked with 0x1f before the shift is applied.
1368 dst.x = src0.x << (0x1f \& src1.x)
1370 dst.y = src0.y << (0x1f \& src1.y)
1372 dst.z = src0.z << (0x1f \& src1.z)
1374 dst.w = src0.w << (0x1f \& src1.w)
1377 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1379 The shift count is masked with 0x1f before the shift is applied.
1383 dst.x = src0.x >> (0x1f \& src1.x)
1385 dst.y = src0.y >> (0x1f \& src1.y)
1387 dst.z = src0.z >> (0x1f \& src1.z)
1389 dst.w = src0.w >> (0x1f \& src1.w)
1392 .. opcode:: USHR - Logical Shift Right
1394 The shift count is masked with 0x1f before the shift is applied.
1398 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1400 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1402 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1404 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1407 .. opcode:: UCMP - Integer Conditional Move
1411 dst.x = src0.x ? src1.x : src2.x
1413 dst.y = src0.y ? src1.y : src2.y
1415 dst.z = src0.z ? src1.z : src2.z
1417 dst.w = src0.w ? src1.w : src2.w
1421 .. opcode:: ISSG - Integer Set Sign
1425 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1427 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1429 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1431 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1435 .. opcode:: FSLT - Float Set On Less Than (ordered)
1437 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1441 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1443 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1445 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1447 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1450 .. opcode:: ISLT - Signed Integer Set On Less Than
1454 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1456 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1458 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1460 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1463 .. opcode:: USLT - Unsigned Integer Set On Less Than
1467 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1469 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1471 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1473 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1476 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1478 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1482 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1484 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1486 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1488 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1491 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1495 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1497 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1499 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1501 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1504 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1508 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1510 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1512 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1514 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1517 .. opcode:: FSEQ - Float Set On Equal (ordered)
1519 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1523 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1525 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1527 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1529 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1532 .. opcode:: USEQ - Integer Set On Equal
1536 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1538 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1540 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1542 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1545 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1547 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1551 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1553 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1555 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1557 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1560 .. opcode:: USNE - Integer Set On Not Equal
1564 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1566 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1568 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1570 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1573 .. opcode:: INEG - Integer Negate
1588 .. opcode:: IABS - Integer Absolute Value
1602 These opcodes are used for bit-level manipulation of integers.
1604 .. opcode:: IBFE - Signed Bitfield Extract
1606 Like GLSL bitfieldExtract. Extracts a set of bits from the input, and
1607 sign-extends them if the high bit of the extracted window is set.
1611 def ibfe(value, offset, bits):
1612 if offset < 0 or bits < 0 or offset + bits > 32:
1614 if bits == 0: return 0
1615 # Note: >> sign-extends
1616 return (value << (32 - offset - bits)) >> (32 - bits)
1618 .. opcode:: UBFE - Unsigned Bitfield Extract
1620 Like GLSL bitfieldExtract. Extracts a set of bits from the input, without
1625 def ubfe(value, offset, bits):
1626 if offset < 0 or bits < 0 or offset + bits > 32:
1628 if bits == 0: return 0
1629 # Note: >> does not sign-extend
1630 return (value << (32 - offset - bits)) >> (32 - bits)
1632 .. opcode:: BFI - Bitfield Insert
1634 Like GLSL bitfieldInsert. Replaces a bit region of 'base' with the low bits
1639 def bfi(base, insert, offset, bits):
1640 if offset < 0 or bits < 0 or offset + bits > 32:
1642 # << defined such that mask == ~0 when bits == 32, offset == 0
1643 mask = ((1 << bits) - 1) << offset
1644 return ((insert << offset) & mask) | (base & ~mask)
1646 .. opcode:: BREV - Bitfield Reverse
1648 See SM5 instruction BFREV. Reverses the bits of the argument.
1650 .. opcode:: POPC - Population Count
1652 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1654 .. opcode:: LSB - Index of lowest set bit
1656 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1657 bit of the argument. Returns -1 if none are set.
1659 .. opcode:: IMSB - Index of highest non-sign bit
1661 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1662 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1663 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1664 (i.e. for inputs 0 and -1).
1666 .. opcode:: UMSB - Index of highest set bit
1668 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1669 set bit of the argument. Returns -1 if none are set.
1672 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1674 These opcodes are only supported in geometry shaders; they have no meaning
1675 in any other type of shader.
1677 .. opcode:: EMIT - Emit
1679 Generate a new vertex for the current primitive into the specified vertex
1680 stream using the values in the output registers.
1683 .. opcode:: ENDPRIM - End Primitive
1685 Complete the current primitive in the specified vertex stream (consisting of
1686 the emitted vertices), and start a new one.
1692 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1693 opcodes is determined by a special capability bit, ``GLSL``.
1694 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1696 .. opcode:: CAL - Subroutine Call
1702 .. opcode:: RET - Subroutine Call Return
1707 .. opcode:: CONT - Continue
1709 Unconditionally moves the point of execution to the instruction after the
1710 last bgnloop. The instruction must appear within a bgnloop/endloop.
1714 Support for CONT is determined by a special capability bit,
1715 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1718 .. opcode:: BGNLOOP - Begin a Loop
1720 Start a loop. Must have a matching endloop.
1723 .. opcode:: BGNSUB - Begin Subroutine
1725 Starts definition of a subroutine. Must have a matching endsub.
1728 .. opcode:: ENDLOOP - End a Loop
1730 End a loop started with bgnloop.
1733 .. opcode:: ENDSUB - End Subroutine
1735 Ends definition of a subroutine.
1738 .. opcode:: NOP - No Operation
1743 .. opcode:: BRK - Break
1745 Unconditionally moves the point of execution to the instruction after the
1746 next endloop or endswitch. The instruction must appear within a loop/endloop
1747 or switch/endswitch.
1750 .. opcode:: BREAKC - Break Conditional
1752 Conditionally moves the point of execution to the instruction after the
1753 next endloop or endswitch. The instruction must appear within a loop/endloop
1754 or switch/endswitch.
1755 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1756 as an integer register.
1760 Considered for removal as it's quite inconsistent wrt other opcodes
1761 (could emulate with UIF/BRK/ENDIF).
1764 .. opcode:: IF - Float If
1766 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1770 where src0.x is interpreted as a floating point register.
1773 .. opcode:: UIF - Bitwise If
1775 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1779 where src0.x is interpreted as an integer register.
1782 .. opcode:: ELSE - Else
1784 Starts an else block, after an IF or UIF statement.
1787 .. opcode:: ENDIF - End If
1789 Ends an IF or UIF block.
1792 .. opcode:: SWITCH - Switch
1794 Starts a C-style switch expression. The switch consists of one or multiple
1795 CASE statements, and at most one DEFAULT statement. Execution of a statement
1796 ends when a BRK is hit, but just like in C falling through to other cases
1797 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1798 just as last statement, and fallthrough is allowed into/from it.
1799 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1805 (some instructions here)
1808 (some instructions here)
1811 (some instructions here)
1816 .. opcode:: CASE - Switch case
1818 This represents a switch case label. The src arg must be an integer immediate.
1821 .. opcode:: DEFAULT - Switch default
1823 This represents the default case in the switch, which is taken if no other
1827 .. opcode:: ENDSWITCH - End of switch
1829 Ends a switch expression.
1835 The interpolation instructions allow an input to be interpolated in a
1836 different way than its declaration. This corresponds to the GLSL 4.00
1837 interpolateAt* functions. The first argument of each of these must come from
1838 ``TGSI_FILE_INPUT``.
1840 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1842 Interpolates the varying specified by src0 at the centroid
1844 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1846 Interpolates the varying specified by src0 at the sample id specified by
1847 src1.x (interpreted as an integer)
1849 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1851 Interpolates the varying specified by src0 at the offset src1.xy from the
1852 pixel center (interpreted as floats)
1860 The double-precision opcodes reinterpret four-component vectors into
1861 two-component vectors with doubled precision in each component.
1863 .. opcode:: DABS - Absolute
1871 .. opcode:: DADD - Add
1875 dst.xy = src0.xy + src1.xy
1877 dst.zw = src0.zw + src1.zw
1879 .. opcode:: DSEQ - Set on Equal
1883 dst.x = src0.xy == src1.xy ? \sim 0 : 0
1885 dst.z = src0.zw == src1.zw ? \sim 0 : 0
1887 .. opcode:: DSNE - Set on Equal
1891 dst.x = src0.xy != src1.xy ? \sim 0 : 0
1893 dst.z = src0.zw != src1.zw ? \sim 0 : 0
1895 .. opcode:: DSLT - Set on Less than
1899 dst.x = src0.xy < src1.xy ? \sim 0 : 0
1901 dst.z = src0.zw < src1.zw ? \sim 0 : 0
1903 .. opcode:: DSGE - Set on Greater equal
1907 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
1909 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
1911 .. opcode:: DFRAC - Fraction
1915 dst.xy = src.xy - \lfloor src.xy\rfloor
1917 dst.zw = src.zw - \lfloor src.zw\rfloor
1919 .. opcode:: DTRUNC - Truncate
1923 dst.xy = trunc(src.xy)
1925 dst.zw = trunc(src.zw)
1927 .. opcode:: DCEIL - Ceiling
1931 dst.xy = \lceil src.xy\rceil
1933 dst.zw = \lceil src.zw\rceil
1935 .. opcode:: DFLR - Floor
1939 dst.xy = \lfloor src.xy\rfloor
1941 dst.zw = \lfloor src.zw\rfloor
1943 .. opcode:: DROUND - Fraction
1947 dst.xy = round(src.xy)
1949 dst.zw = round(src.zw)
1951 .. opcode:: DSSG - Set Sign
1955 dst.xy = (src.xy > 0) ? 1.0 : (src.xy < 0) ? -1.0 : 0.0
1957 dst.zw = (src.zw > 0) ? 1.0 : (src.zw < 0) ? -1.0 : 0.0
1959 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1961 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1962 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1963 :math:`dst1 \times 2^{dst0} = src` .
1967 dst0.xy = exp(src.xy)
1969 dst1.xy = frac(src.xy)
1971 dst0.zw = exp(src.zw)
1973 dst1.zw = frac(src.zw)
1975 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1977 This opcode is the inverse of :opcode:`DFRACEXP`. The second
1978 source is an integer.
1982 dst.xy = src0.xy \times 2^{src1.x}
1984 dst.zw = src0.zw \times 2^{src1.y}
1986 .. opcode:: DMIN - Minimum
1990 dst.xy = min(src0.xy, src1.xy)
1992 dst.zw = min(src0.zw, src1.zw)
1994 .. opcode:: DMAX - Maximum
1998 dst.xy = max(src0.xy, src1.xy)
2000 dst.zw = max(src0.zw, src1.zw)
2002 .. opcode:: DMUL - Multiply
2006 dst.xy = src0.xy \times src1.xy
2008 dst.zw = src0.zw \times src1.zw
2011 .. opcode:: DMAD - Multiply And Add
2015 dst.xy = src0.xy \times src1.xy + src2.xy
2017 dst.zw = src0.zw \times src1.zw + src2.zw
2020 .. opcode:: DFMA - Fused Multiply-Add
2022 Perform a * b + c with no intermediate rounding step.
2026 dst.xy = src0.xy \times src1.xy + src2.xy
2028 dst.zw = src0.zw \times src1.zw + src2.zw
2031 .. opcode:: DDIV - Divide
2035 dst.xy = \frac{src0.xy}{src1.xy}
2037 dst.zw = \frac{src0.zw}{src1.zw}
2040 .. opcode:: DRCP - Reciprocal
2044 dst.xy = \frac{1}{src.xy}
2046 dst.zw = \frac{1}{src.zw}
2048 .. opcode:: DSQRT - Square Root
2052 dst.xy = \sqrt{src.xy}
2054 dst.zw = \sqrt{src.zw}
2056 .. opcode:: DRSQ - Reciprocal Square Root
2060 dst.xy = \frac{1}{\sqrt{src.xy}}
2062 dst.zw = \frac{1}{\sqrt{src.zw}}
2064 .. opcode:: F2D - Float to Double
2068 dst.xy = double(src0.x)
2070 dst.zw = double(src0.y)
2072 .. opcode:: D2F - Double to Float
2076 dst.x = float(src0.xy)
2078 dst.y = float(src0.zw)
2080 .. opcode:: I2D - Int to Double
2084 dst.xy = double(src0.x)
2086 dst.zw = double(src0.y)
2088 .. opcode:: D2I - Double to Int
2092 dst.x = int(src0.xy)
2094 dst.y = int(src0.zw)
2096 .. opcode:: U2D - Unsigned Int to Double
2100 dst.xy = double(src0.x)
2102 dst.zw = double(src0.y)
2104 .. opcode:: D2U - Double to Unsigned Int
2108 dst.x = unsigned(src0.xy)
2110 dst.y = unsigned(src0.zw)
2115 The 64-bit integer opcodes reinterpret four-component vectors into
2116 two-component vectors with 64-bits in each component.
2118 .. opcode:: I64ABS - 64-bit Integer Absolute Value
2126 .. opcode:: I64NEG - 64-bit Integer Negate
2136 .. opcode:: I64SSG - 64-bit Integer Set Sign
2140 dst.xy = (src0.xy < 0) ? -1 : (src0.xy > 0) ? 1 : 0
2142 dst.zw = (src0.zw < 0) ? -1 : (src0.zw > 0) ? 1 : 0
2144 .. opcode:: U64ADD - 64-bit Integer Add
2148 dst.xy = src0.xy + src1.xy
2150 dst.zw = src0.zw + src1.zw
2152 .. opcode:: U64MUL - 64-bit Integer Multiply
2156 dst.xy = src0.xy * src1.xy
2158 dst.zw = src0.zw * src1.zw
2160 .. opcode:: U64SEQ - 64-bit Integer Set on Equal
2164 dst.x = src0.xy == src1.xy ? \sim 0 : 0
2166 dst.z = src0.zw == src1.zw ? \sim 0 : 0
2168 .. opcode:: U64SNE - 64-bit Integer Set on Not Equal
2172 dst.x = src0.xy != src1.xy ? \sim 0 : 0
2174 dst.z = src0.zw != src1.zw ? \sim 0 : 0
2176 .. opcode:: U64SLT - 64-bit Unsigned Integer Set on Less Than
2180 dst.x = src0.xy < src1.xy ? \sim 0 : 0
2182 dst.z = src0.zw < src1.zw ? \sim 0 : 0
2184 .. opcode:: U64SGE - 64-bit Unsigned Integer Set on Greater Equal
2188 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
2190 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
2192 .. opcode:: I64SLT - 64-bit Signed Integer Set on Less Than
2196 dst.x = src0.xy < src1.xy ? \sim 0 : 0
2198 dst.z = src0.zw < src1.zw ? \sim 0 : 0
2200 .. opcode:: I64SGE - 64-bit Signed Integer Set on Greater Equal
2204 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
2206 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
2208 .. opcode:: I64MIN - Minimum of 64-bit Signed Integers
2212 dst.xy = min(src0.xy, src1.xy)
2214 dst.zw = min(src0.zw, src1.zw)
2216 .. opcode:: U64MIN - Minimum of 64-bit Unsigned Integers
2220 dst.xy = min(src0.xy, src1.xy)
2222 dst.zw = min(src0.zw, src1.zw)
2224 .. opcode:: I64MAX - Maximum of 64-bit Signed Integers
2228 dst.xy = max(src0.xy, src1.xy)
2230 dst.zw = max(src0.zw, src1.zw)
2232 .. opcode:: U64MAX - Maximum of 64-bit Unsigned Integers
2236 dst.xy = max(src0.xy, src1.xy)
2238 dst.zw = max(src0.zw, src1.zw)
2240 .. opcode:: U64SHL - Shift Left 64-bit Unsigned Integer
2242 The shift count is masked with 0x3f before the shift is applied.
2246 dst.xy = src0.xy << (0x3f \& src1.x)
2248 dst.zw = src0.zw << (0x3f \& src1.y)
2250 .. opcode:: I64SHR - Arithmetic Shift Right (of 64-bit Signed Integer)
2252 The shift count is masked with 0x3f before the shift is applied.
2256 dst.xy = src0.xy >> (0x3f \& src1.x)
2258 dst.zw = src0.zw >> (0x3f \& src1.y)
2260 .. opcode:: U64SHR - Logical Shift Right (of 64-bit Unsigned Integer)
2262 The shift count is masked with 0x3f before the shift is applied.
2266 dst.xy = src0.xy >> (unsigned) (0x3f \& src1.x)
2268 dst.zw = src0.zw >> (unsigned) (0x3f \& src1.y)
2270 .. opcode:: I64DIV - 64-bit Signed Integer Division
2274 dst.xy = \frac{src0.xy}{src1.xy}
2276 dst.zw = \frac{src0.zw}{src1.zw}
2278 .. opcode:: U64DIV - 64-bit Unsigned Integer Division
2282 dst.xy = \frac{src0.xy}{src1.xy}
2284 dst.zw = \frac{src0.zw}{src1.zw}
2286 .. opcode:: U64MOD - 64-bit Unsigned Integer Remainder
2290 dst.xy = src0.xy \bmod src1.xy
2292 dst.zw = src0.zw \bmod src1.zw
2294 .. opcode:: I64MOD - 64-bit Signed Integer Remainder
2298 dst.xy = src0.xy \bmod src1.xy
2300 dst.zw = src0.zw \bmod src1.zw
2302 .. opcode:: F2U64 - Float to 64-bit Unsigned Int
2306 dst.xy = (uint64_t) src0.x
2308 dst.zw = (uint64_t) src0.y
2310 .. opcode:: F2I64 - Float to 64-bit Int
2314 dst.xy = (int64_t) src0.x
2316 dst.zw = (int64_t) src0.y
2318 .. opcode:: U2I64 - Unsigned Integer to 64-bit Integer
2320 This is a zero extension.
2324 dst.xy = (uint64_t) src0.x
2326 dst.zw = (uint64_t) src0.y
2328 .. opcode:: I2I64 - Signed Integer to 64-bit Integer
2330 This is a sign extension.
2334 dst.xy = (int64_t) src0.x
2336 dst.zw = (int64_t) src0.y
2338 .. opcode:: D2U64 - Double to 64-bit Unsigned Int
2342 dst.xy = (uint64_t) src0.xy
2344 dst.zw = (uint64_t) src0.zw
2346 .. opcode:: D2I64 - Double to 64-bit Int
2350 dst.xy = (int64_t) src0.xy
2352 dst.zw = (int64_t) src0.zw
2354 .. opcode:: U642F - 64-bit unsigned integer to float
2358 dst.x = (float) src0.xy
2360 dst.y = (float) src0.zw
2362 .. opcode:: I642F - 64-bit Int to Float
2366 dst.x = (float) src0.xy
2368 dst.y = (float) src0.zw
2370 .. opcode:: U642D - 64-bit unsigned integer to double
2374 dst.xy = (double) src0.xy
2376 dst.zw = (double) src0.zw
2378 .. opcode:: I642D - 64-bit Int to double
2382 dst.xy = (double) src0.xy
2384 dst.zw = (double) src0.zw
2386 .. _samplingopcodes:
2388 Resource Sampling Opcodes
2389 ^^^^^^^^^^^^^^^^^^^^^^^^^
2391 Those opcodes follow very closely semantics of the respective Direct3D
2392 instructions. If in doubt double check Direct3D documentation.
2393 Note that the swizzle on SVIEW (src1) determines texel swizzling
2398 Using provided address, sample data from the specified texture using the
2399 filtering mode identified by the given sampler. The source data may come from
2400 any resource type other than buffers.
2402 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
2404 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
2406 .. opcode:: SAMPLE_I
2408 Simplified alternative to the SAMPLE instruction. Using the provided
2409 integer address, SAMPLE_I fetches data from the specified sampler view
2410 without any filtering. The source data may come from any resource type
2413 Syntax: ``SAMPLE_I dst, address, sampler_view``
2415 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
2417 The 'address' is specified as unsigned integers. If the 'address' is out of
2418 range [0...(# texels - 1)] the result of the fetch is always 0 in all
2419 components. As such the instruction doesn't honor address wrap modes, in
2420 cases where that behavior is desirable 'SAMPLE' instruction should be used.
2421 address.w always provides an unsigned integer mipmap level. If the value is
2422 out of the range then the instruction always returns 0 in all components.
2423 address.yz are ignored for buffers and 1d textures. address.z is ignored
2424 for 1d texture arrays and 2d textures.
2426 For 1D texture arrays address.y provides the array index (also as unsigned
2427 integer). If the value is out of the range of available array indices
2428 [0... (array size - 1)] then the opcode always returns 0 in all components.
2429 For 2D texture arrays address.z provides the array index, otherwise it
2430 exhibits the same behavior as in the case for 1D texture arrays. The exact
2431 semantics of the source address are presented in the table below:
2433 +---------------------------+----+-----+-----+---------+
2434 | resource type | X | Y | Z | W |
2435 +===========================+====+=====+=====+=========+
2436 | ``PIPE_BUFFER`` | x | | | ignored |
2437 +---------------------------+----+-----+-----+---------+
2438 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2439 +---------------------------+----+-----+-----+---------+
2440 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2441 +---------------------------+----+-----+-----+---------+
2442 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2443 +---------------------------+----+-----+-----+---------+
2444 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2445 +---------------------------+----+-----+-----+---------+
2446 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2447 +---------------------------+----+-----+-----+---------+
2448 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2449 +---------------------------+----+-----+-----+---------+
2450 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2451 +---------------------------+----+-----+-----+---------+
2453 Where 'mpl' is a mipmap level and 'idx' is the array index.
2455 .. opcode:: SAMPLE_I_MS
2457 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2459 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2461 .. opcode:: SAMPLE_B
2463 Just like the SAMPLE instruction with the exception that an additional bias
2464 is applied to the level of detail computed as part of the instruction
2467 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2469 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2471 .. opcode:: SAMPLE_C
2473 Similar to the SAMPLE instruction but it performs a comparison filter. The
2474 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2475 additional float32 operand, reference value, which must be a register with
2476 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2477 current samplers compare_func (in pipe_sampler_state) to compare reference
2478 value against the red component value for the surce resource at each texel
2479 that the currently configured texture filter covers based on the provided
2482 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2484 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2486 .. opcode:: SAMPLE_C_LZ
2488 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2491 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2493 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2496 .. opcode:: SAMPLE_D
2498 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2499 the source address in the x direction and the y direction are provided by
2502 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2504 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2506 .. opcode:: SAMPLE_L
2508 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2509 directly as a scalar value, representing no anisotropy.
2511 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2513 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2517 Gathers the four texels to be used in a bi-linear filtering operation and
2518 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2519 and cubemaps arrays. For 2D textures, only the addressing modes of the
2520 sampler and the top level of any mip pyramid are used. Set W to zero. It
2521 behaves like the SAMPLE instruction, but a filtered sample is not
2522 generated. The four samples that contribute to filtering are placed into
2523 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2524 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2525 magnitude of the deltas are half a texel.
2528 .. opcode:: SVIEWINFO
2530 Query the dimensions of a given sampler view. dst receives width, height,
2531 depth or array size and number of mipmap levels as int4. The dst can have a
2532 writemask which will specify what info is the caller interested in.
2534 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2536 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2538 src_mip_level is an unsigned integer scalar. If it's out of range then
2539 returns 0 for width, height and depth/array size but the total number of
2540 mipmap is still returned correctly for the given sampler view. The returned
2541 width, height and depth values are for the mipmap level selected by the
2542 src_mip_level and are in the number of texels. For 1d texture array width
2543 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2544 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2545 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2546 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2547 resinfo allowing swizzling dst values is ignored (due to the interaction
2548 with rcpfloat modifier which requires some swizzle handling in the state
2551 .. opcode:: SAMPLE_POS
2553 Query the position of a sample in the given resource or render target
2554 when per-sample fragment shading is in effect.
2556 Syntax: ``SAMPLE_POS dst, source, sample_index``
2558 dst receives float4 (x, y, undef, undef) indicated where the sample is
2559 located. Sample locations are in the range [0, 1] where 0.5 is the center
2562 source is either a sampler view (to indicate a shader resource) or temp
2563 register (to indicate the render target). The source register may have
2564 an optional swizzle to apply to the returned result
2566 sample_index is an integer scalar indicating which sample position is to
2569 If per-sample shading is not in effect or the source resource or render
2570 target is not multisampled, the result is (0.5, 0.5, undef, undef).
2572 NOTE: no driver has implemented this opcode yet (and no state tracker
2573 emits it). This information is subject to change.
2575 .. opcode:: SAMPLE_INFO
2577 Query the number of samples in a multisampled resource or render target.
2579 Syntax: ``SAMPLE_INFO dst, source``
2581 dst receives int4 (n, 0, 0, 0) where n is the number of samples in a
2582 resource or the render target.
2584 source is either a sampler view (to indicate a shader resource) or temp
2585 register (to indicate the render target). The source register may have
2586 an optional swizzle to apply to the returned result
2588 If per-sample shading is not in effect or the source resource or render
2589 target is not multisampled, the result is (1, 0, 0, 0).
2591 NOTE: no driver has implemented this opcode yet (and no state tracker
2592 emits it). This information is subject to change.
2594 .. _resourceopcodes:
2596 Resource Access Opcodes
2597 ^^^^^^^^^^^^^^^^^^^^^^^
2599 For these opcodes, the resource can be a BUFFER, IMAGE, or MEMORY.
2601 .. opcode:: LOAD - Fetch data from a shader buffer or image
2603 Syntax: ``LOAD dst, resource, address``
2605 Example: ``LOAD TEMP[0], BUFFER[0], TEMP[1]``
2607 Using the provided integer address, LOAD fetches data
2608 from the specified buffer or texture without any
2611 The 'address' is specified as a vector of unsigned
2612 integers. If the 'address' is out of range the result
2615 Only the first mipmap level of a resource can be read
2616 from using this instruction.
2618 For 1D or 2D texture arrays, the array index is
2619 provided as an unsigned integer in address.y or
2620 address.z, respectively. address.yz are ignored for
2621 buffers and 1D textures. address.z is ignored for 1D
2622 texture arrays and 2D textures. address.w is always
2625 A swizzle suffix may be added to the resource argument
2626 this will cause the resource data to be swizzled accordingly.
2628 .. opcode:: STORE - Write data to a shader resource
2630 Syntax: ``STORE resource, address, src``
2632 Example: ``STORE BUFFER[0], TEMP[0], TEMP[1]``
2634 Using the provided integer address, STORE writes data
2635 to the specified buffer or texture.
2637 The 'address' is specified as a vector of unsigned
2638 integers. If the 'address' is out of range the result
2641 Only the first mipmap level of a resource can be
2642 written to using this instruction.
2644 For 1D or 2D texture arrays, the array index is
2645 provided as an unsigned integer in address.y or
2646 address.z, respectively. address.yz are ignored for
2647 buffers and 1D textures. address.z is ignored for 1D
2648 texture arrays and 2D textures. address.w is always
2651 .. opcode:: RESQ - Query information about a resource
2653 Syntax: ``RESQ dst, resource``
2655 Example: ``RESQ TEMP[0], BUFFER[0]``
2657 Returns information about the buffer or image resource. For buffer
2658 resources, the size (in bytes) is returned in the x component. For
2659 image resources, .xyz will contain the width/height/layers of the
2660 image, while .w will contain the number of samples for multi-sampled
2663 .. opcode:: FBFETCH - Load data from framebuffer
2665 Syntax: ``FBFETCH dst, output``
2667 Example: ``FBFETCH TEMP[0], OUT[0]``
2669 This is only valid on ``COLOR`` semantic outputs. Returns the color
2670 of the current position in the framebuffer from before this fragment
2671 shader invocation. May return the same value from multiple calls for
2672 a particular output within a single invocation. Note that result may
2673 be undefined if a fragment is drawn multiple times without a blend
2677 .. _threadsyncopcodes:
2679 Inter-thread synchronization opcodes
2680 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2682 These opcodes are intended for communication between threads running
2683 within the same compute grid. For now they're only valid in compute
2686 .. opcode:: MFENCE - Memory fence
2688 Syntax: ``MFENCE resource``
2690 Example: ``MFENCE RES[0]``
2692 This opcode forces strong ordering between any memory access
2693 operations that affect the specified resource. This means that
2694 previous loads and stores (and only those) will be performed and
2695 visible to other threads before the program execution continues.
2698 .. opcode:: LFENCE - Load memory fence
2700 Syntax: ``LFENCE resource``
2702 Example: ``LFENCE RES[0]``
2704 Similar to MFENCE, but it only affects the ordering of memory loads.
2707 .. opcode:: SFENCE - Store memory fence
2709 Syntax: ``SFENCE resource``
2711 Example: ``SFENCE RES[0]``
2713 Similar to MFENCE, but it only affects the ordering of memory stores.
2716 .. opcode:: BARRIER - Thread group barrier
2720 This opcode suspends the execution of the current thread until all
2721 the remaining threads in the working group reach the same point of
2722 the program. Results are unspecified if any of the remaining
2723 threads terminates or never reaches an executed BARRIER instruction.
2725 .. opcode:: MEMBAR - Memory barrier
2729 This opcode waits for the completion of all memory accesses based on
2730 the type passed in. The type is an immediate bitfield with the following
2733 Bit 0: Shader storage buffers
2734 Bit 1: Atomic buffers
2736 Bit 3: Shared memory
2739 These may be passed in in any combination. An implementation is free to not
2740 distinguish between these as it sees fit. However these map to all the
2741 possibilities made available by GLSL.
2748 These opcodes provide atomic variants of some common arithmetic and
2749 logical operations. In this context atomicity means that another
2750 concurrent memory access operation that affects the same memory
2751 location is guaranteed to be performed strictly before or after the
2752 entire execution of the atomic operation. The resource may be a BUFFER,
2753 IMAGE, or MEMORY. In the case of an image, the offset works the same as for
2754 ``LOAD`` and ``STORE``, specified above. These atomic operations may
2755 only be used with 32-bit integer image formats.
2757 .. opcode:: ATOMUADD - Atomic integer addition
2759 Syntax: ``ATOMUADD dst, resource, offset, src``
2761 Example: ``ATOMUADD TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2763 The following operation is performed atomically:
2767 dst_x = resource[offset]
2769 resource[offset] = dst_x + src_x
2772 .. opcode:: ATOMXCHG - Atomic exchange
2774 Syntax: ``ATOMXCHG dst, resource, offset, src``
2776 Example: ``ATOMXCHG TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2778 The following operation is performed atomically:
2782 dst_x = resource[offset]
2784 resource[offset] = src_x
2787 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2789 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2791 Example: ``ATOMCAS TEMP[0], BUFFER[0], TEMP[1], TEMP[2], TEMP[3]``
2793 The following operation is performed atomically:
2797 dst_x = resource[offset]
2799 resource[offset] = (dst_x == cmp_x ? src_x : dst_x)
2802 .. opcode:: ATOMAND - Atomic bitwise And
2804 Syntax: ``ATOMAND dst, resource, offset, src``
2806 Example: ``ATOMAND TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2808 The following operation is performed atomically:
2812 dst_x = resource[offset]
2814 resource[offset] = dst_x \& src_x
2817 .. opcode:: ATOMOR - Atomic bitwise Or
2819 Syntax: ``ATOMOR dst, resource, offset, src``
2821 Example: ``ATOMOR TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2823 The following operation is performed atomically:
2827 dst_x = resource[offset]
2829 resource[offset] = dst_x | src_x
2832 .. opcode:: ATOMXOR - Atomic bitwise Xor
2834 Syntax: ``ATOMXOR dst, resource, offset, src``
2836 Example: ``ATOMXOR TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2838 The following operation is performed atomically:
2842 dst_x = resource[offset]
2844 resource[offset] = dst_x \oplus src_x
2847 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2849 Syntax: ``ATOMUMIN dst, resource, offset, src``
2851 Example: ``ATOMUMIN TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2853 The following operation is performed atomically:
2857 dst_x = resource[offset]
2859 resource[offset] = (dst_x < src_x ? dst_x : src_x)
2862 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2864 Syntax: ``ATOMUMAX dst, resource, offset, src``
2866 Example: ``ATOMUMAX TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2868 The following operation is performed atomically:
2872 dst_x = resource[offset]
2874 resource[offset] = (dst_x > src_x ? dst_x : src_x)
2877 .. opcode:: ATOMIMIN - Atomic signed minimum
2879 Syntax: ``ATOMIMIN dst, resource, offset, src``
2881 Example: ``ATOMIMIN TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2883 The following operation is performed atomically:
2887 dst_x = resource[offset]
2889 resource[offset] = (dst_x < src_x ? dst_x : src_x)
2892 .. opcode:: ATOMIMAX - Atomic signed maximum
2894 Syntax: ``ATOMIMAX dst, resource, offset, src``
2896 Example: ``ATOMIMAX TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2898 The following operation is performed atomically:
2902 dst_x = resource[offset]
2904 resource[offset] = (dst_x > src_x ? dst_x : src_x)
2907 .. _interlaneopcodes:
2912 These opcodes reduce the given value across the shader invocations
2913 running in the current SIMD group. Every thread in the subgroup will receive
2914 the same result. The BALLOT operations accept a single-channel argument that
2915 is treated as a boolean and produce a 64-bit value.
2917 .. opcode:: VOTE_ANY - Value is set in any of the active invocations
2919 Syntax: ``VOTE_ANY dst, value``
2921 Example: ``VOTE_ANY TEMP[0].x, TEMP[1].x``
2924 .. opcode:: VOTE_ALL - Value is set in all of the active invocations
2926 Syntax: ``VOTE_ALL dst, value``
2928 Example: ``VOTE_ALL TEMP[0].x, TEMP[1].x``
2931 .. opcode:: VOTE_EQ - Value is the same in all of the active invocations
2933 Syntax: ``VOTE_EQ dst, value``
2935 Example: ``VOTE_EQ TEMP[0].x, TEMP[1].x``
2938 .. opcode:: BALLOT - Lanemask of whether the value is set in each active
2941 Syntax: ``BALLOT dst, value``
2943 Example: ``BALLOT TEMP[0].xy, TEMP[1].x``
2945 When the argument is a constant true, this produces a bitmask of active
2946 invocations. In fragment shaders, this can include helper invocations
2947 (invocations whose outputs and writes to memory are discarded, but which
2948 are used to compute derivatives).
2951 .. opcode:: READ_FIRST - Broadcast the value from the first active
2952 invocation to all active lanes
2954 Syntax: ``READ_FIRST dst, value``
2956 Example: ``READ_FIRST TEMP[0], TEMP[1]``
2959 .. opcode:: READ_INVOC - Retrieve the value from the given invocation
2960 (need not be uniform)
2962 Syntax: ``READ_INVOC dst, value, invocation``
2964 Example: ``READ_INVOC TEMP[0].xy, TEMP[1].xy, TEMP[2].x``
2966 invocation.x controls the invocation number to read from for all channels.
2967 The invocation number must be the same across all active invocations in a
2968 sub-group; otherwise, the results are undefined.
2971 Explanation of symbols used
2972 ------------------------------
2979 :math:`|x|` Absolute value of `x`.
2981 :math:`\lceil x \rceil` Ceiling of `x`.
2983 clamp(x,y,z) Clamp x between y and z.
2984 (x < y) ? y : (x > z) ? z : x
2986 :math:`\lfloor x\rfloor` Floor of `x`.
2988 :math:`\log_2{x}` Logarithm of `x`, base 2.
2990 max(x,y) Maximum of x and y.
2993 min(x,y) Minimum of x and y.
2996 partialx(x) Derivative of x relative to fragment's X.
2998 partialy(x) Derivative of x relative to fragment's Y.
3000 pop() Pop from stack.
3002 :math:`x^y` `x` to the power `y`.
3004 push(x) Push x on stack.
3008 trunc(x) Truncate x, i.e. drop the fraction bits.
3015 discard Discard fragment.
3019 target Label of target instruction.
3030 Declares a register that is will be referenced as an operand in Instruction
3033 File field contains register file that is being declared and is one
3036 UsageMask field specifies which of the register components can be accessed
3037 and is one of TGSI_WRITEMASK.
3039 The Local flag specifies that a given value isn't intended for
3040 subroutine parameter passing and, as a result, the implementation
3041 isn't required to give any guarantees of it being preserved across
3042 subroutine boundaries. As it's merely a compiler hint, the
3043 implementation is free to ignore it.
3045 If Dimension flag is set to 1, a Declaration Dimension token follows.
3047 If Semantic flag is set to 1, a Declaration Semantic token follows.
3049 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
3051 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
3053 If Array flag is set to 1, a Declaration Array token follows.
3056 ^^^^^^^^^^^^^^^^^^^^^^^^
3058 Declarations can optional have an ArrayID attribute which can be referred by
3059 indirect addressing operands. An ArrayID of zero is reserved and treated as
3060 if no ArrayID is specified.
3062 If an indirect addressing operand refers to a specific declaration by using
3063 an ArrayID only the registers in this declaration are guaranteed to be
3064 accessed, accessing any register outside this declaration results in undefined
3065 behavior. Note that for compatibility the effective index is zero-based and
3066 not relative to the specified declaration
3068 If no ArrayID is specified with an indirect addressing operand the whole
3069 register file might be accessed by this operand. This is strongly discouraged
3070 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
3071 This is only legal for TEMP and CONST register files.
3073 Declaration Semantic
3074 ^^^^^^^^^^^^^^^^^^^^^^^^
3076 Vertex and fragment shader input and output registers may be labeled
3077 with semantic information consisting of a name and index.
3079 Follows Declaration token if Semantic bit is set.
3081 Since its purpose is to link a shader with other stages of the pipeline,
3082 it is valid to follow only those Declaration tokens that declare a register
3083 either in INPUT or OUTPUT file.
3085 SemanticName field contains the semantic name of the register being declared.
3086 There is no default value.
3088 SemanticIndex is an optional subscript that can be used to distinguish
3089 different register declarations with the same semantic name. The default value
3092 The meanings of the individual semantic names are explained in the following
3095 TGSI_SEMANTIC_POSITION
3096 """"""""""""""""""""""
3098 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
3099 output register which contains the homogeneous vertex position in the clip
3100 space coordinate system. After clipping, the X, Y and Z components of the
3101 vertex will be divided by the W value to get normalized device coordinates.
3103 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
3104 fragment shader input (or system value, depending on which one is
3105 supported by the driver) contains the fragment's window position. The X
3106 component starts at zero and always increases from left to right.
3107 The Y component starts at zero and always increases but Y=0 may either
3108 indicate the top of the window or the bottom depending on the fragment
3109 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
3110 The Z coordinate ranges from 0 to 1 to represent depth from the front
3111 to the back of the Z buffer. The W component contains the interpolated
3112 reciprocal of the vertex position W component (corresponding to gl_Fragcoord,
3113 but unlike d3d10 which interpolates the same 1/w but then gives back
3114 the reciprocal of the interpolated value).
3116 Fragment shaders may also declare an output register with
3117 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
3118 the fragment shader to change the fragment's Z position.
3125 For vertex shader outputs or fragment shader inputs/outputs, this
3126 label indicates that the register contains an R,G,B,A color.
3128 Several shader inputs/outputs may contain colors so the semantic index
3129 is used to distinguish them. For example, color[0] may be the diffuse
3130 color while color[1] may be the specular color.
3132 This label is needed so that the flat/smooth shading can be applied
3133 to the right interpolants during rasterization.
3137 TGSI_SEMANTIC_BCOLOR
3138 """"""""""""""""""""
3140 Back-facing colors are only used for back-facing polygons, and are only valid
3141 in vertex shader outputs. After rasterization, all polygons are front-facing
3142 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
3143 so all BCOLORs effectively become regular COLORs in the fragment shader.
3149 Vertex shader inputs and outputs and fragment shader inputs may be
3150 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
3151 a fog coordinate. Typically, the fragment shader will use the fog coordinate
3152 to compute a fog blend factor which is used to blend the normal fragment color
3153 with a constant fog color. But fog coord really is just an ordinary vec4
3154 register like regular semantics.
3160 Vertex shader input and output registers may be labeled with
3161 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
3162 in the form (S, 0, 0, 1). The point size controls the width or diameter
3163 of points for rasterization. This label cannot be used in fragment
3166 When using this semantic, be sure to set the appropriate state in the
3167 :ref:`rasterizer` first.
3170 TGSI_SEMANTIC_TEXCOORD
3171 """"""""""""""""""""""
3173 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
3175 Vertex shader outputs and fragment shader inputs may be labeled with
3176 this semantic to make them replaceable by sprite coordinates via the
3177 sprite_coord_enable state in the :ref:`rasterizer`.
3178 The semantic index permitted with this semantic is limited to <= 7.
3180 If the driver does not support TEXCOORD, sprite coordinate replacement
3181 applies to inputs with the GENERIC semantic instead.
3183 The intended use case for this semantic is gl_TexCoord.
3186 TGSI_SEMANTIC_PCOORD
3187 """"""""""""""""""""
3189 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
3191 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
3192 that the register contains sprite coordinates in the form (x, y, 0, 1), if
3193 the current primitive is a point and point sprites are enabled. Otherwise,
3194 the contents of the register are undefined.
3196 The intended use case for this semantic is gl_PointCoord.
3199 TGSI_SEMANTIC_GENERIC
3200 """""""""""""""""""""
3202 All vertex/fragment shader inputs/outputs not labeled with any other
3203 semantic label can be considered to be generic attributes. Typical
3204 uses of generic inputs/outputs are texcoords and user-defined values.
3207 TGSI_SEMANTIC_NORMAL
3208 """"""""""""""""""""
3210 Indicates that a vertex shader input is a normal vector. This is
3211 typically only used for legacy graphics APIs.
3217 This label applies to fragment shader inputs (or system values,
3218 depending on which one is supported by the driver) and indicates that
3219 the register contains front/back-face information.
3221 If it is an input, it will be a floating-point vector in the form (F, 0, 0, 1),
3222 where F will be positive when the fragment belongs to a front-facing polygon,
3223 and negative when the fragment belongs to a back-facing polygon.
3225 If it is a system value, it will be an integer vector in the form (F, 0, 0, 1),
3226 where F is 0xffffffff when the fragment belongs to a front-facing polygon and
3227 0 when the fragment belongs to a back-facing polygon.
3230 TGSI_SEMANTIC_EDGEFLAG
3231 """"""""""""""""""""""
3233 For vertex shaders, this sematic label indicates that an input or
3234 output is a boolean edge flag. The register layout is [F, x, x, x]
3235 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
3236 simply copies the edge flag input to the edgeflag output.
3238 Edge flags are used to control which lines or points are actually
3239 drawn when the polygon mode converts triangles/quads/polygons into
3243 TGSI_SEMANTIC_STENCIL
3244 """""""""""""""""""""
3246 For fragment shaders, this semantic label indicates that an output
3247 is a writable stencil reference value. Only the Y component is writable.
3248 This allows the fragment shader to change the fragments stencilref value.
3251 TGSI_SEMANTIC_VIEWPORT_INDEX
3252 """"""""""""""""""""""""""""
3254 For geometry shaders, this semantic label indicates that an output
3255 contains the index of the viewport (and scissor) to use.
3256 This is an integer value, and only the X component is used.
3258 If PIPE_CAP_TGSI_VS_LAYER_VIEWPORT or PIPE_CAP_TGSI_TES_LAYER_VIEWPORT is
3259 supported, then this semantic label can also be used in vertex or
3260 tessellation evaluation shaders, respectively. Only the value written in the
3261 last vertex processing stage is used.
3267 For geometry shaders, this semantic label indicates that an output
3268 contains the layer value to use for the color and depth/stencil surfaces.
3269 This is an integer value, and only the X component is used.
3270 (Also known as rendertarget array index.)
3272 If PIPE_CAP_TGSI_VS_LAYER_VIEWPORT or PIPE_CAP_TGSI_TES_LAYER_VIEWPORT is
3273 supported, then this semantic label can also be used in vertex or
3274 tessellation evaluation shaders, respectively. Only the value written in the
3275 last vertex processing stage is used.
3278 TGSI_SEMANTIC_CULLDIST
3279 """"""""""""""""""""""
3281 Used as distance to plane for performing application-defined culling
3282 of individual primitives against a plane. When components of vertex
3283 elements are given this label, these values are assumed to be a
3284 float32 signed distance to a plane. Primitives will be completely
3285 discarded if the plane distance for all of the vertices in the
3286 primitive are < 0. If a vertex has a cull distance of NaN, that
3287 vertex counts as "out" (as if its < 0);
3288 The limits on both clip and cull distances are bound
3289 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
3290 the maximum number of components that can be used to hold the
3291 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
3292 which specifies the maximum number of registers which can be
3293 annotated with those semantics.
3296 TGSI_SEMANTIC_CLIPDIST
3297 """"""""""""""""""""""
3299 Note this covers clipping and culling distances.
3301 When components of vertex elements are identified this way, these
3302 values are each assumed to be a float32 signed distance to a plane.
3305 Primitive setup only invokes rasterization on pixels for which
3306 the interpolated plane distances are >= 0.
3309 Primitives will be completely discarded if the plane distance
3310 for all of the vertices in the primitive are < 0.
3311 If a vertex has a cull distance of NaN, that vertex counts as "out"
3314 Multiple clip/cull planes can be implemented simultaneously, by
3315 annotating multiple components of one or more vertex elements with
3316 the above specified semantic.
3317 The limits on both clip and cull distances are bound
3318 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
3319 the maximum number of components that can be used to hold the
3320 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
3321 which specifies the maximum number of registers which can be
3322 annotated with those semantics.
3323 The properties NUM_CLIPDIST_ENABLED and NUM_CULLDIST_ENABLED
3324 are used to divide up the 2 x vec4 space between clipping and culling.
3326 TGSI_SEMANTIC_SAMPLEID
3327 """"""""""""""""""""""
3329 For fragment shaders, this semantic label indicates that a system value
3330 contains the current sample id (i.e. gl_SampleID) as an unsigned int.
3331 Only the X component is used. If per-sample shading is not enabled,
3332 the result is (0, undef, undef, undef).
3334 Note that if the fragment shader uses this system value, the fragment
3335 shader is automatically executed at per sample frequency.
3337 TGSI_SEMANTIC_SAMPLEPOS
3338 """""""""""""""""""""""
3340 For fragment shaders, this semantic label indicates that a system
3341 value contains the current sample's position as float4(x, y, undef, undef)
3342 in the render target (i.e. gl_SamplePosition) when per-fragment shading
3343 is in effect. Position values are in the range [0, 1] where 0.5 is
3344 the center of the fragment.
3346 Note that if the fragment shader uses this system value, the fragment
3347 shader is automatically executed at per sample frequency.
3349 TGSI_SEMANTIC_SAMPLEMASK
3350 """"""""""""""""""""""""
3352 For fragment shaders, this semantic label can be applied to either a
3353 shader system value input or output.
3355 For a system value, the sample mask indicates the set of samples covered by
3356 the current primitive. If MSAA is not enabled, the value is (1, 0, 0, 0).
3358 For an output, the sample mask is used to disable further sample processing.
3360 For both, the register type is uint[4] but only the X component is used
3361 (i.e. gl_SampleMask[0]). Each bit corresponds to one sample position (up
3362 to 32x MSAA is supported).
3364 TGSI_SEMANTIC_INVOCATIONID
3365 """"""""""""""""""""""""""
3367 For geometry shaders, this semantic label indicates that a system value
3368 contains the current invocation id (i.e. gl_InvocationID).
3369 This is an integer value, and only the X component is used.
3371 TGSI_SEMANTIC_INSTANCEID
3372 """"""""""""""""""""""""
3374 For vertex shaders, this semantic label indicates that a system value contains
3375 the current instance id (i.e. gl_InstanceID). It does not include the base
3376 instance. This is an integer value, and only the X component is used.
3378 TGSI_SEMANTIC_VERTEXID
3379 """"""""""""""""""""""
3381 For vertex shaders, this semantic label indicates that a system value contains
3382 the current vertex id (i.e. gl_VertexID). It does (unlike in d3d10) include the
3383 base vertex. This is an integer value, and only the X component is used.
3385 TGSI_SEMANTIC_VERTEXID_NOBASE
3386 """""""""""""""""""""""""""""""
3388 For vertex shaders, this semantic label indicates that a system value contains
3389 the current vertex id without including the base vertex (this corresponds to
3390 d3d10 vertex id, so TGSI_SEMANTIC_VERTEXID_NOBASE + TGSI_SEMANTIC_BASEVERTEX
3391 == TGSI_SEMANTIC_VERTEXID). This is an integer value, and only the X component
3394 TGSI_SEMANTIC_BASEVERTEX
3395 """"""""""""""""""""""""
3397 For vertex shaders, this semantic label indicates that a system value contains
3398 the base vertex (i.e. gl_BaseVertex). Note that for non-indexed draw calls,
3399 this contains the first (or start) value instead.
3400 This is an integer value, and only the X component is used.
3402 TGSI_SEMANTIC_PRIMID
3403 """"""""""""""""""""
3405 For geometry and fragment shaders, this semantic label indicates the value
3406 contains the primitive id (i.e. gl_PrimitiveID). This is an integer value,
3407 and only the X component is used.
3408 FIXME: This right now can be either a ordinary input or a system value...
3414 For tessellation evaluation/control shaders, this semantic label indicates a
3415 generic per-patch attribute. Such semantics will not implicitly be per-vertex
3418 TGSI_SEMANTIC_TESSCOORD
3419 """""""""""""""""""""""
3421 For tessellation evaluation shaders, this semantic label indicates the
3422 coordinates of the vertex being processed. This is available in XYZ; W is
3425 TGSI_SEMANTIC_TESSOUTER
3426 """""""""""""""""""""""
3428 For tessellation evaluation/control shaders, this semantic label indicates the
3429 outer tessellation levels of the patch. Isoline tessellation will only have XY
3430 defined, triangle will have XYZ and quads will have XYZW defined. This
3431 corresponds to gl_TessLevelOuter.
3433 TGSI_SEMANTIC_TESSINNER
3434 """""""""""""""""""""""
3436 For tessellation evaluation/control shaders, this semantic label indicates the
3437 inner tessellation levels of the patch. The X value is only defined for
3438 triangle tessellation, while quads will have XY defined. This is entirely
3439 undefined for isoline tessellation.
3441 TGSI_SEMANTIC_VERTICESIN
3442 """"""""""""""""""""""""
3444 For tessellation evaluation/control shaders, this semantic label indicates the
3445 number of vertices provided in the input patch. Only the X value is defined.
3447 TGSI_SEMANTIC_HELPER_INVOCATION
3448 """""""""""""""""""""""""""""""
3450 For fragment shaders, this semantic indicates whether the current
3451 invocation is covered or not. Helper invocations are created in order
3452 to properly compute derivatives, however it may be desirable to skip
3453 some of the logic in those cases. See ``gl_HelperInvocation`` documentation.
3455 TGSI_SEMANTIC_BASEINSTANCE
3456 """"""""""""""""""""""""""
3458 For vertex shaders, the base instance argument supplied for this
3459 draw. This is an integer value, and only the X component is used.
3461 TGSI_SEMANTIC_DRAWID
3462 """"""""""""""""""""
3464 For vertex shaders, the zero-based index of the current draw in a
3465 ``glMultiDraw*`` invocation. This is an integer value, and only the X
3469 TGSI_SEMANTIC_WORK_DIM
3470 """"""""""""""""""""""
3472 For compute shaders started via opencl this retrieves the work_dim
3473 parameter to the clEnqueueNDRangeKernel call with which the shader
3477 TGSI_SEMANTIC_GRID_SIZE
3478 """""""""""""""""""""""
3480 For compute shaders, this semantic indicates the maximum (x, y, z) dimensions
3481 of a grid of thread blocks.
3484 TGSI_SEMANTIC_BLOCK_ID
3485 """"""""""""""""""""""
3487 For compute shaders, this semantic indicates the (x, y, z) coordinates of the
3488 current block inside of the grid.
3491 TGSI_SEMANTIC_BLOCK_SIZE
3492 """"""""""""""""""""""""
3494 For compute shaders, this semantic indicates the maximum (x, y, z) dimensions
3495 of a block in threads.
3498 TGSI_SEMANTIC_THREAD_ID
3499 """""""""""""""""""""""
3501 For compute shaders, this semantic indicates the (x, y, z) coordinates of the
3502 current thread inside of the block.
3505 TGSI_SEMANTIC_SUBGROUP_SIZE
3506 """""""""""""""""""""""""""
3508 This semantic indicates the subgroup size for the current invocation. This is
3509 an integer of at most 64, as it indicates the width of lanemasks. It does not
3510 depend on the number of invocations that are active.
3513 TGSI_SEMANTIC_SUBGROUP_INVOCATION
3514 """""""""""""""""""""""""""""""""
3516 The index of the current invocation within its subgroup.
3519 TGSI_SEMANTIC_SUBGROUP_EQ_MASK
3520 """"""""""""""""""""""""""""""
3522 A bit mask of ``bit index == TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3523 ``1 << subgroup_invocation`` in arbitrary precision arithmetic.
3526 TGSI_SEMANTIC_SUBGROUP_GE_MASK
3527 """"""""""""""""""""""""""""""
3529 A bit mask of ``bit index >= TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3530 ``((1 << (subgroup_size - subgroup_invocation)) - 1) << subgroup_invocation``
3531 in arbitrary precision arithmetic.
3534 TGSI_SEMANTIC_SUBGROUP_GT_MASK
3535 """"""""""""""""""""""""""""""
3537 A bit mask of ``bit index > TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3538 ``((1 << (subgroup_size - subgroup_invocation - 1)) - 1) << (subgroup_invocation + 1)``
3539 in arbitrary precision arithmetic.
3542 TGSI_SEMANTIC_SUBGROUP_LE_MASK
3543 """"""""""""""""""""""""""""""
3545 A bit mask of ``bit index <= TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3546 ``(1 << (subgroup_invocation + 1)) - 1`` in arbitrary precision arithmetic.
3549 TGSI_SEMANTIC_SUBGROUP_LT_MASK
3550 """"""""""""""""""""""""""""""
3552 A bit mask of ``bit index > TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3553 ``(1 << subgroup_invocation) - 1`` in arbitrary precision arithmetic.
3556 Declaration Interpolate
3557 ^^^^^^^^^^^^^^^^^^^^^^^
3559 This token is only valid for fragment shader INPUT declarations.
3561 The Interpolate field specifes the way input is being interpolated by
3562 the rasteriser and is one of TGSI_INTERPOLATE_*.
3564 The Location field specifies the location inside the pixel that the
3565 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
3566 when per-sample shading is enabled, the implementation may choose to
3567 interpolate at the sample irrespective of the Location field.
3569 The CylindricalWrap bitfield specifies which register components
3570 should be subject to cylindrical wrapping when interpolating by the
3571 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
3572 should be interpolated according to cylindrical wrapping rules.
3575 Declaration Sampler View
3576 ^^^^^^^^^^^^^^^^^^^^^^^^
3578 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
3580 DCL SVIEW[#], resource, type(s)
3582 Declares a shader input sampler view and assigns it to a SVIEW[#]
3585 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
3587 type must be 1 or 4 entries (if specifying on a per-component
3588 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
3590 For TEX\* style texture sample opcodes (as opposed to SAMPLE\* opcodes
3591 which take an explicit SVIEW[#] source register), there may be optionally
3592 SVIEW[#] declarations. In this case, the SVIEW index is implied by the
3593 SAMP index, and there must be a corresponding SVIEW[#] declaration for
3594 each SAMP[#] declaration. Drivers are free to ignore this if they wish.
3595 But note in particular that some drivers need to know the sampler type
3596 (float/int/unsigned) in order to generate the correct code, so cases
3597 where integer textures are sampled, SVIEW[#] declarations should be
3600 NOTE: It is NOT legal to mix SAMPLE\* style opcodes and TEX\* opcodes
3603 Declaration Resource
3604 ^^^^^^^^^^^^^^^^^^^^
3606 Follows Declaration token if file is TGSI_FILE_RESOURCE.
3608 DCL RES[#], resource [, WR] [, RAW]
3610 Declares a shader input resource and assigns it to a RES[#]
3613 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
3616 If the RAW keyword is not specified, the texture data will be
3617 subject to conversion, swizzling and scaling as required to yield
3618 the specified data type from the physical data format of the bound
3621 If the RAW keyword is specified, no channel conversion will be
3622 performed: the values read for each of the channels (X,Y,Z,W) will
3623 correspond to consecutive words in the same order and format
3624 they're found in memory. No element-to-address conversion will be
3625 performed either: the value of the provided X coordinate will be
3626 interpreted in byte units instead of texel units. The result of
3627 accessing a misaligned address is undefined.
3629 Usage of the STORE opcode is only allowed if the WR (writable) flag
3634 ^^^^^^^^^^^^^^^^^^^^^^^^
3636 Properties are general directives that apply to the whole TGSI program.
3641 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
3642 The default value is UPPER_LEFT.
3644 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
3645 increase downward and rightward.
3646 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
3647 increase upward and rightward.
3649 OpenGL defaults to LOWER_LEFT, and is configurable with the
3650 GL_ARB_fragment_coord_conventions extension.
3652 DirectX 9/10 use UPPER_LEFT.
3654 FS_COORD_PIXEL_CENTER
3655 """""""""""""""""""""
3657 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
3658 The default value is HALF_INTEGER.
3660 If HALF_INTEGER, the fractionary part of the position will be 0.5
3661 If INTEGER, the fractionary part of the position will be 0.0
3663 Note that this does not affect the set of fragments generated by
3664 rasterization, which is instead controlled by half_pixel_center in the
3667 OpenGL defaults to HALF_INTEGER, and is configurable with the
3668 GL_ARB_fragment_coord_conventions extension.
3670 DirectX 9 uses INTEGER.
3671 DirectX 10 uses HALF_INTEGER.
3673 FS_COLOR0_WRITES_ALL_CBUFS
3674 """"""""""""""""""""""""""
3675 Specifies that writes to the fragment shader color 0 are replicated to all
3676 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
3677 fragData is directed to a single color buffer, but fragColor is broadcast.
3680 """"""""""""""""""""""""""
3681 If this property is set on the program bound to the shader stage before the
3682 fragment shader, user clip planes should have no effect (be disabled) even if
3683 that shader does not write to any clip distance outputs and the rasterizer's
3684 clip_plane_enable is non-zero.
3685 This property is only supported by drivers that also support shader clip
3687 This is useful for APIs that don't have UCPs and where clip distances written
3688 by a shader cannot be disabled.
3693 Specifies the number of times a geometry shader should be executed for each
3694 input primitive. Each invocation will have a different
3695 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
3698 VS_WINDOW_SPACE_POSITION
3699 """"""""""""""""""""""""""
3700 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
3701 is assumed to contain window space coordinates.
3702 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
3703 directly taken from the 4-th component of the shader output.
3704 Naturally, clipping is not performed on window coordinates either.
3705 The effect of this property is undefined if a geometry or tessellation shader
3711 The number of vertices written by the tessellation control shader. This
3712 effectively defines the patch input size of the tessellation evaluation shader
3718 This sets the tessellation primitive mode, one of ``PIPE_PRIM_TRIANGLES``,
3719 ``PIPE_PRIM_QUADS``, or ``PIPE_PRIM_LINES``. (Unlike in GL, there is no
3720 separate isolines settings, the regular lines is assumed to mean isolines.)
3725 This sets the spacing mode of the tessellation generator, one of
3726 ``PIPE_TESS_SPACING_*``.
3731 This sets the vertex order to be clockwise if the value is 1, or
3732 counter-clockwise if set to 0.
3737 If set to a non-zero value, this turns on point mode for the tessellator,
3738 which means that points will be generated instead of primitives.
3740 NUM_CLIPDIST_ENABLED
3741 """"""""""""""""""""
3743 How many clip distance scalar outputs are enabled.
3745 NUM_CULLDIST_ENABLED
3746 """"""""""""""""""""
3748 How many cull distance scalar outputs are enabled.
3750 FS_EARLY_DEPTH_STENCIL
3751 """"""""""""""""""""""
3753 Whether depth test, stencil test, and occlusion query should run before
3754 the fragment shader (regardless of fragment shader side effects). Corresponds
3755 to GLSL early_fragment_tests.
3760 Which shader stage will MOST LIKELY follow after this shader when the shader
3761 is bound. This is only a hint to the driver and doesn't have to be precise.
3762 Only set for VS and TES.
3764 CS_FIXED_BLOCK_WIDTH / HEIGHT / DEPTH
3765 """""""""""""""""""""""""""""""""""""
3767 Threads per block in each dimension, if known at compile time. If the block size
3768 is known all three should be at least 1. If it is unknown they should all be set
3774 The MUL TGSI operation (FP32 multiplication) will return 0 if either
3775 of the operands are equal to 0. That means that 0 * Inf = 0. This
3776 should be set the same way for an entire pipeline. Note that this
3777 applies not only to the literal MUL TGSI opcode, but all FP32
3778 multiplications implied by other operations, such as MAD, FMA, DP2,
3779 DP3, DP4, DPH, DST, LOG, LRP, XPD, and possibly others. If there is a
3780 mismatch between shaders, then it is unspecified whether this behavior
3783 FS_POST_DEPTH_COVERAGE
3784 """"""""""""""""""""""
3786 When enabled, the input for TGSI_SEMANTIC_SAMPLEMASK will exclude samples
3787 that have failed the depth/stencil tests. This is only valid when
3788 FS_EARLY_DEPTH_STENCIL is also specified.
3791 Texture Sampling and Texture Formats
3792 ------------------------------------
3794 This table shows how texture image components are returned as (x,y,z,w) tuples
3795 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
3796 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
3799 +--------------------+--------------+--------------------+--------------+
3800 | Texture Components | Gallium | OpenGL | Direct3D 9 |
3801 +====================+==============+====================+==============+
3802 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
3803 +--------------------+--------------+--------------------+--------------+
3804 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
3805 +--------------------+--------------+--------------------+--------------+
3806 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
3807 +--------------------+--------------+--------------------+--------------+
3808 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
3809 +--------------------+--------------+--------------------+--------------+
3810 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
3811 +--------------------+--------------+--------------------+--------------+
3812 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
3813 +--------------------+--------------+--------------------+--------------+
3814 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
3815 +--------------------+--------------+--------------------+--------------+
3816 | I | (i, i, i, i) | (i, i, i, i) | N/A |
3817 +--------------------+--------------+--------------------+--------------+
3818 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
3819 | | | [#envmap-bumpmap]_ | |
3820 +--------------------+--------------+--------------------+--------------+
3821 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
3822 | | | [#depth-tex-mode]_ | |
3823 +--------------------+--------------+--------------------+--------------+
3824 | S | (s, s, s, s) | unknown | unknown |
3825 +--------------------+--------------+--------------------+--------------+
3827 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
3828 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
3829 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.