4 TGSI, Tungsten Graphics Shader Infrastructure, is an intermediate language
5 for describing shaders. Since Gallium is inherently shaderful, shaders are
6 an important part of the API. TGSI is the only intermediate representation
12 All TGSI instructions, known as *opcodes*, operate on arbitrary-precision
13 floating-point four-component vectors. An opcode may have up to one
14 destination register, known as *dst*, and between zero and three source
15 registers, called *src0* through *src2*, or simply *src* if there is only
18 Some instructions, like :opcode:`I2F`, permit re-interpretation of vector
19 components as integers. Other instructions permit using registers as
20 two-component vectors with double precision; see :ref:`doubleopcodes`.
22 When an instruction has a scalar result, the result is usually copied into
23 each of the components of *dst*. When this happens, the result is said to be
24 *replicated* to *dst*. :opcode:`RCP` is one such instruction.
29 TGSI supports modifiers on inputs (as well as saturate modifier on instructions).
31 For inputs which have a floating point type, both absolute value and
32 negation modifiers are supported (with absolute value being applied
33 first). The only source of TGSI_OPCODE_MOV and the second and third
34 sources of TGSI_OPCODE_UCMP are considered to have float type for
37 For inputs which have signed or unsigned type only the negate modifier is
44 ^^^^^^^^^^^^^^^^^^^^^^^^^
46 These opcodes are guaranteed to be available regardless of the driver being
49 .. opcode:: ARL - Address Register Load
53 dst.x = (int) \lfloor src.x\rfloor
55 dst.y = (int) \lfloor src.y\rfloor
57 dst.z = (int) \lfloor src.z\rfloor
59 dst.w = (int) \lfloor src.w\rfloor
62 .. opcode:: MOV - Move
75 .. opcode:: LIT - Light Coefficients
80 dst.y &= max(src.x, 0) \\
81 dst.z &= (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0 \\
85 .. opcode:: RCP - Reciprocal
87 This instruction replicates its result.
94 .. opcode:: RSQ - Reciprocal Square Root
96 This instruction replicates its result. The results are undefined for src <= 0.
100 dst = \frac{1}{\sqrt{src.x}}
103 .. opcode:: SQRT - Square Root
105 This instruction replicates its result. The results are undefined for src < 0.
112 .. opcode:: EXP - Approximate Exponential Base 2
116 dst.x &= 2^{\lfloor src.x\rfloor} \\
117 dst.y &= src.x - \lfloor src.x\rfloor \\
118 dst.z &= 2^{src.x} \\
122 .. opcode:: LOG - Approximate Logarithm Base 2
126 dst.x &= \lfloor\log_2{|src.x|}\rfloor \\
127 dst.y &= \frac{|src.x|}{2^{\lfloor\log_2{|src.x|}\rfloor}} \\
128 dst.z &= \log_2{|src.x|} \\
132 .. opcode:: MUL - Multiply
136 dst.x = src0.x \times src1.x
138 dst.y = src0.y \times src1.y
140 dst.z = src0.z \times src1.z
142 dst.w = src0.w \times src1.w
145 .. opcode:: ADD - Add
149 dst.x = src0.x + src1.x
151 dst.y = src0.y + src1.y
153 dst.z = src0.z + src1.z
155 dst.w = src0.w + src1.w
158 .. opcode:: DP3 - 3-component Dot Product
160 This instruction replicates its result.
164 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z
167 .. opcode:: DP4 - 4-component Dot Product
169 This instruction replicates its result.
173 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src0.w \times src1.w
176 .. opcode:: DST - Distance Vector
181 dst.y &= src0.y \times src1.y\\
186 .. opcode:: MIN - Minimum
190 dst.x = min(src0.x, src1.x)
192 dst.y = min(src0.y, src1.y)
194 dst.z = min(src0.z, src1.z)
196 dst.w = min(src0.w, src1.w)
199 .. opcode:: MAX - Maximum
203 dst.x = max(src0.x, src1.x)
205 dst.y = max(src0.y, src1.y)
207 dst.z = max(src0.z, src1.z)
209 dst.w = max(src0.w, src1.w)
212 .. opcode:: SLT - Set On Less Than
216 dst.x = (src0.x < src1.x) ? 1.0F : 0.0F
218 dst.y = (src0.y < src1.y) ? 1.0F : 0.0F
220 dst.z = (src0.z < src1.z) ? 1.0F : 0.0F
222 dst.w = (src0.w < src1.w) ? 1.0F : 0.0F
225 .. opcode:: SGE - Set On Greater Equal Than
229 dst.x = (src0.x >= src1.x) ? 1.0F : 0.0F
231 dst.y = (src0.y >= src1.y) ? 1.0F : 0.0F
233 dst.z = (src0.z >= src1.z) ? 1.0F : 0.0F
235 dst.w = (src0.w >= src1.w) ? 1.0F : 0.0F
238 .. opcode:: MAD - Multiply And Add
242 dst.x = src0.x \times src1.x + src2.x
244 dst.y = src0.y \times src1.y + src2.y
246 dst.z = src0.z \times src1.z + src2.z
248 dst.w = src0.w \times src1.w + src2.w
251 .. opcode:: LRP - Linear Interpolate
255 dst.x = src0.x \times src1.x + (1 - src0.x) \times src2.x
257 dst.y = src0.y \times src1.y + (1 - src0.y) \times src2.y
259 dst.z = src0.z \times src1.z + (1 - src0.z) \times src2.z
261 dst.w = src0.w \times src1.w + (1 - src0.w) \times src2.w
264 .. opcode:: FMA - Fused Multiply-Add
266 Perform a * b + c with no intermediate rounding step.
270 dst.x = src0.x \times src1.x + src2.x
272 dst.y = src0.y \times src1.y + src2.y
274 dst.z = src0.z \times src1.z + src2.z
276 dst.w = src0.w \times src1.w + src2.w
279 .. opcode:: DP2A - 2-component Dot Product And Add
283 dst.x = src0.x \times src1.x + src0.y \times src1.y + src2.x
285 dst.y = src0.x \times src1.x + src0.y \times src1.y + src2.x
287 dst.z = src0.x \times src1.x + src0.y \times src1.y + src2.x
289 dst.w = src0.x \times src1.x + src0.y \times src1.y + src2.x
292 .. opcode:: FRC - Fraction
296 dst.x = src.x - \lfloor src.x\rfloor
298 dst.y = src.y - \lfloor src.y\rfloor
300 dst.z = src.z - \lfloor src.z\rfloor
302 dst.w = src.w - \lfloor src.w\rfloor
305 .. opcode:: FLR - Floor
309 dst.x = \lfloor src.x\rfloor
311 dst.y = \lfloor src.y\rfloor
313 dst.z = \lfloor src.z\rfloor
315 dst.w = \lfloor src.w\rfloor
318 .. opcode:: ROUND - Round
331 .. opcode:: EX2 - Exponential Base 2
333 This instruction replicates its result.
340 .. opcode:: LG2 - Logarithm Base 2
342 This instruction replicates its result.
349 .. opcode:: POW - Power
351 This instruction replicates its result.
355 dst = src0.x^{src1.x}
357 .. opcode:: XPD - Cross Product
361 dst.x = src0.y \times src1.z - src1.y \times src0.z
363 dst.y = src0.z \times src1.x - src1.z \times src0.x
365 dst.z = src0.x \times src1.y - src1.x \times src0.y
370 .. opcode:: DPH - Homogeneous Dot Product
372 This instruction replicates its result.
376 dst = src0.x \times src1.x + src0.y \times src1.y + src0.z \times src1.z + src1.w
379 .. opcode:: COS - Cosine
381 This instruction replicates its result.
388 .. opcode:: DDX, DDX_FINE - Derivative Relative To X
390 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
391 advertised. When it is, the fine version guarantees one derivative per row
392 while DDX is allowed to be the same for the entire 2x2 quad.
396 dst.x = partialx(src.x)
398 dst.y = partialx(src.y)
400 dst.z = partialx(src.z)
402 dst.w = partialx(src.w)
405 .. opcode:: DDY, DDY_FINE - Derivative Relative To Y
407 The fine variant is only used when ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE`` is
408 advertised. When it is, the fine version guarantees one derivative per column
409 while DDY is allowed to be the same for the entire 2x2 quad.
413 dst.x = partialy(src.x)
415 dst.y = partialy(src.y)
417 dst.z = partialy(src.z)
419 dst.w = partialy(src.w)
422 .. opcode:: PK2H - Pack Two 16-bit Floats
424 This instruction replicates its result.
428 dst = f32\_to\_f16(src.x) | f32\_to\_f16(src.y) << 16
431 .. opcode:: PK2US - Pack Two Unsigned 16-bit Scalars
436 .. opcode:: PK4B - Pack Four Signed 8-bit Scalars
441 .. opcode:: PK4UB - Pack Four Unsigned 8-bit Scalars
446 .. opcode:: SEQ - Set On Equal
450 dst.x = (src0.x == src1.x) ? 1.0F : 0.0F
452 dst.y = (src0.y == src1.y) ? 1.0F : 0.0F
454 dst.z = (src0.z == src1.z) ? 1.0F : 0.0F
456 dst.w = (src0.w == src1.w) ? 1.0F : 0.0F
459 .. opcode:: SGT - Set On Greater Than
463 dst.x = (src0.x > src1.x) ? 1.0F : 0.0F
465 dst.y = (src0.y > src1.y) ? 1.0F : 0.0F
467 dst.z = (src0.z > src1.z) ? 1.0F : 0.0F
469 dst.w = (src0.w > src1.w) ? 1.0F : 0.0F
472 .. opcode:: SIN - Sine
474 This instruction replicates its result.
481 .. opcode:: SLE - Set On Less Equal Than
485 dst.x = (src0.x <= src1.x) ? 1.0F : 0.0F
487 dst.y = (src0.y <= src1.y) ? 1.0F : 0.0F
489 dst.z = (src0.z <= src1.z) ? 1.0F : 0.0F
491 dst.w = (src0.w <= src1.w) ? 1.0F : 0.0F
494 .. opcode:: SNE - Set On Not Equal
498 dst.x = (src0.x != src1.x) ? 1.0F : 0.0F
500 dst.y = (src0.y != src1.y) ? 1.0F : 0.0F
502 dst.z = (src0.z != src1.z) ? 1.0F : 0.0F
504 dst.w = (src0.w != src1.w) ? 1.0F : 0.0F
507 .. opcode:: TEX - Texture Lookup
509 for array textures src0.y contains the slice for 1D,
510 and src0.z contain the slice for 2D.
512 for shadow textures with no arrays (and not cube map),
513 src0.z contains the reference value.
515 for shadow textures with arrays, src0.z contains
516 the reference value for 1D arrays, and src0.w contains
517 the reference value for 2D arrays and cube maps.
519 for cube map array shadow textures, the reference value
520 cannot be passed in src0.w, and TEX2 must be used instead.
526 shadow_ref = src0.z or src0.w (optional)
530 dst = texture\_sample(unit, coord, shadow_ref)
533 .. opcode:: TEX2 - Texture Lookup (for shadow cube map arrays only)
535 this is the same as TEX, but uses another reg to encode the
546 dst = texture\_sample(unit, coord, shadow_ref)
551 .. opcode:: TXD - Texture Lookup with Derivatives
563 dst = texture\_sample\_deriv(unit, coord, ddx, ddy)
566 .. opcode:: TXP - Projective Texture Lookup
570 coord.x = src0.x / src0.w
572 coord.y = src0.y / src0.w
574 coord.z = src0.z / src0.w
580 dst = texture\_sample(unit, coord)
583 .. opcode:: UP2H - Unpack Two 16-Bit Floats
587 dst.x = f16\_to\_f32(src0.x \& 0xffff)
589 dst.y = f16\_to\_f32(src0.x >> 16)
591 dst.z = f16\_to\_f32(src0.x \& 0xffff)
593 dst.w = f16\_to\_f32(src0.x >> 16)
597 Considered for removal.
599 .. opcode:: UP2US - Unpack Two Unsigned 16-Bit Scalars
605 Considered for removal.
607 .. opcode:: UP4B - Unpack Four Signed 8-Bit Values
613 Considered for removal.
615 .. opcode:: UP4UB - Unpack Four Unsigned 8-Bit Scalars
621 Considered for removal.
624 .. opcode:: ARR - Address Register Load With Round
628 dst.x = (int) round(src.x)
630 dst.y = (int) round(src.y)
632 dst.z = (int) round(src.z)
634 dst.w = (int) round(src.w)
637 .. opcode:: SSG - Set Sign
641 dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
643 dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
645 dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
647 dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
650 .. opcode:: CMP - Compare
654 dst.x = (src0.x < 0) ? src1.x : src2.x
656 dst.y = (src0.y < 0) ? src1.y : src2.y
658 dst.z = (src0.z < 0) ? src1.z : src2.z
660 dst.w = (src0.w < 0) ? src1.w : src2.w
663 .. opcode:: KILL_IF - Conditional Discard
665 Conditional discard. Allowed in fragment shaders only.
669 if (src.x < 0 || src.y < 0 || src.z < 0 || src.w < 0)
674 .. opcode:: KILL - Discard
676 Unconditional discard. Allowed in fragment shaders only.
679 .. opcode:: SCS - Sine Cosine
692 .. opcode:: TXB - Texture Lookup With Bias
694 for cube map array textures and shadow cube maps, the bias value
695 cannot be passed in src0.w, and TXB2 must be used instead.
697 if the target is a shadow texture, the reference value is always
698 in src.z (this prevents shadow 3d and shadow 2d arrays from
699 using this instruction, but this is not needed).
715 dst = texture\_sample(unit, coord, bias)
718 .. opcode:: TXB2 - Texture Lookup With Bias (some cube maps only)
720 this is the same as TXB, but uses another reg to encode the
721 lod bias value for cube map arrays and shadow cube maps.
722 Presumably shadow 2d arrays and shadow 3d targets could use
723 this encoding too, but this is not legal.
725 shadow cube map arrays are neither possible nor required.
735 dst = texture\_sample(unit, coord, bias)
738 .. opcode:: DIV - Divide
742 dst.x = \frac{src0.x}{src1.x}
744 dst.y = \frac{src0.y}{src1.y}
746 dst.z = \frac{src0.z}{src1.z}
748 dst.w = \frac{src0.w}{src1.w}
751 .. opcode:: DP2 - 2-component Dot Product
753 This instruction replicates its result.
757 dst = src0.x \times src1.x + src0.y \times src1.y
760 .. opcode:: TEX_LZ - Texture Lookup With LOD = 0
762 This is the same as TXL with LOD = 0. Like every texture opcode, it obeys
763 pipe_sampler_view::u.tex.first_level and pipe_sampler_state::min_lod.
764 There is no way to override those two in shaders.
780 dst = texture\_sample(unit, coord, lod)
783 .. opcode:: TXL - Texture Lookup With explicit LOD
785 for cube map array textures, the explicit lod value
786 cannot be passed in src0.w, and TXL2 must be used instead.
788 if the target is a shadow texture, the reference value is always
789 in src.z (this prevents shadow 3d / 2d array / cube targets from
790 using this instruction, but this is not needed).
806 dst = texture\_sample(unit, coord, lod)
809 .. opcode:: TXL2 - Texture Lookup With explicit LOD (for cube map arrays only)
811 this is the same as TXL, but uses another reg to encode the
813 Presumably shadow 3d / 2d array / cube targets could use
814 this encoding too, but this is not legal.
816 shadow cube map arrays are neither possible nor required.
826 dst = texture\_sample(unit, coord, lod)
829 .. opcode:: PUSHA - Push Address Register On Stack
838 Considered for cleanup.
842 Considered for removal.
844 .. opcode:: POPA - Pop Address Register From Stack
853 Considered for cleanup.
857 Considered for removal.
860 .. opcode:: CALLNZ - Subroutine Call If Not Zero
866 Considered for cleanup.
870 Considered for removal.
874 ^^^^^^^^^^^^^^^^^^^^^^^^
876 These opcodes are primarily provided for special-use computational shaders.
877 Support for these opcodes indicated by a special pipe capability bit (TBD).
879 XXX doesn't look like most of the opcodes really belong here.
881 .. opcode:: CEIL - Ceiling
885 dst.x = \lceil src.x\rceil
887 dst.y = \lceil src.y\rceil
889 dst.z = \lceil src.z\rceil
891 dst.w = \lceil src.w\rceil
894 .. opcode:: TRUNC - Truncate
907 .. opcode:: MOD - Modulus
911 dst.x = src0.x \bmod src1.x
913 dst.y = src0.y \bmod src1.y
915 dst.z = src0.z \bmod src1.z
917 dst.w = src0.w \bmod src1.w
920 .. opcode:: UARL - Integer Address Register Load
922 Moves the contents of the source register, assumed to be an integer, into the
923 destination register, which is assumed to be an address (ADDR) register.
926 .. opcode:: SAD - Sum Of Absolute Differences
930 dst.x = |src0.x - src1.x| + src2.x
932 dst.y = |src0.y - src1.y| + src2.y
934 dst.z = |src0.z - src1.z| + src2.z
936 dst.w = |src0.w - src1.w| + src2.w
939 .. opcode:: TXF - Texel Fetch
941 As per NV_gpu_shader4, extract a single texel from a specified texture
942 image or PIPE_BUFFER resource. The source sampler may not be a CUBE or
944 four-component signed integer vector used to identify the single texel
945 accessed. 3 components + level. Just like texture instructions, an optional
946 offset vector is provided, which is subject to various driver restrictions
947 (regarding range, source of offsets). This instruction ignores the sampler
950 TXF(uint_vec coord, int_vec offset).
953 .. opcode:: TXF_LZ - Texel Fetch
955 This is the same as TXF with level = 0. Like TXF, it obeys
956 pipe_sampler_view::u.tex.first_level.
959 .. opcode:: TXQ - Texture Size Query
961 As per NV_gpu_program4, retrieve the dimensions of the texture depending on
962 the target. For 1D (width), 2D/RECT/CUBE (width, height), 3D (width, height,
963 depth), 1D array (width, layers), 2D array (width, height, layers).
964 Also return the number of accessible levels (last_level - first_level + 1)
967 For components which don't return a resource dimension, their value
974 dst.x = texture\_width(unit, lod)
976 dst.y = texture\_height(unit, lod)
978 dst.z = texture\_depth(unit, lod)
980 dst.w = texture\_levels(unit)
983 .. opcode:: TXQS - Texture Samples Query
985 This retrieves the number of samples in the texture, and stores it
986 into the x component as an unsigned integer. The other components are
987 undefined. If the texture is not multisampled, this function returns
988 (1, undef, undef, undef).
992 dst.x = texture\_samples(unit)
995 .. opcode:: TG4 - Texture Gather
997 As per ARB_texture_gather, gathers the four texels to be used in a bi-linear
998 filtering operation and packs them into a single register. Only works with
999 2D, 2D array, cubemaps, and cubemaps arrays. For 2D textures, only the
1000 addressing modes of the sampler and the top level of any mip pyramid are
1001 used. Set W to zero. It behaves like the TEX instruction, but a filtered
1002 sample is not generated. The four samples that contribute to filtering are
1003 placed into xyzw in clockwise order, starting with the (u,v) texture
1004 coordinate delta at the following locations (-, +), (+, +), (+, -), (-, -),
1005 where the magnitude of the deltas are half a texel.
1007 PIPE_CAP_TEXTURE_SM5 enhances this instruction to support shadow per-sample
1008 depth compares, single component selection, and a non-constant offset. It
1009 doesn't allow support for the GL independent offset to get i0,j0. This would
1010 require another CAP is hw can do it natively. For now we lower that before
1019 dst = texture\_gather4 (unit, coord, component)
1021 (with SM5 - cube array shadow)
1029 dst = texture\_gather (uint, coord, compare)
1031 .. opcode:: LODQ - level of detail query
1033 Compute the LOD information that the texture pipe would use to access the
1034 texture. The Y component contains the computed LOD lambda_prime. The X
1035 component contains the LOD that will be accessed, based on min/max lod's
1042 dst.xy = lodq(uint, coord);
1044 .. opcode:: CLOCK - retrieve the current shader time
1046 Invoking this instruction multiple times in the same shader should
1047 cause monotonically increasing values to be returned. The values
1048 are implicitly 64-bit, so if fewer than 64 bits of precision are
1049 available, to provide expected wraparound semantics, the value
1050 should be shifted up so that the most significant bit of the time
1051 is the most significant bit of the 64-bit value.
1059 ^^^^^^^^^^^^^^^^^^^^^^^^
1060 These opcodes are used for integer operations.
1061 Support for these opcodes indicated by PIPE_SHADER_CAP_INTEGERS (all of them?)
1064 .. opcode:: I2F - Signed Integer To Float
1066 Rounding is unspecified (round to nearest even suggested).
1070 dst.x = (float) src.x
1072 dst.y = (float) src.y
1074 dst.z = (float) src.z
1076 dst.w = (float) src.w
1079 .. opcode:: U2F - Unsigned Integer To Float
1081 Rounding is unspecified (round to nearest even suggested).
1085 dst.x = (float) src.x
1087 dst.y = (float) src.y
1089 dst.z = (float) src.z
1091 dst.w = (float) src.w
1094 .. opcode:: F2I - Float to Signed Integer
1096 Rounding is towards zero (truncate).
1097 Values outside signed range (including NaNs) produce undefined results.
1110 .. opcode:: F2U - Float to Unsigned Integer
1112 Rounding is towards zero (truncate).
1113 Values outside unsigned range (including NaNs) produce undefined results.
1117 dst.x = (unsigned) src.x
1119 dst.y = (unsigned) src.y
1121 dst.z = (unsigned) src.z
1123 dst.w = (unsigned) src.w
1126 .. opcode:: UADD - Integer Add
1128 This instruction works the same for signed and unsigned integers.
1129 The low 32bit of the result is returned.
1133 dst.x = src0.x + src1.x
1135 dst.y = src0.y + src1.y
1137 dst.z = src0.z + src1.z
1139 dst.w = src0.w + src1.w
1142 .. opcode:: UMAD - Integer Multiply And Add
1144 This instruction works the same for signed and unsigned integers.
1145 The multiplication returns the low 32bit (as does the result itself).
1149 dst.x = src0.x \times src1.x + src2.x
1151 dst.y = src0.y \times src1.y + src2.y
1153 dst.z = src0.z \times src1.z + src2.z
1155 dst.w = src0.w \times src1.w + src2.w
1158 .. opcode:: UMUL - Integer Multiply
1160 This instruction works the same for signed and unsigned integers.
1161 The low 32bit of the result is returned.
1165 dst.x = src0.x \times src1.x
1167 dst.y = src0.y \times src1.y
1169 dst.z = src0.z \times src1.z
1171 dst.w = src0.w \times src1.w
1174 .. opcode:: IMUL_HI - Signed Integer Multiply High Bits
1176 The high 32bits of the multiplication of 2 signed integers are returned.
1180 dst.x = (src0.x \times src1.x) >> 32
1182 dst.y = (src0.y \times src1.y) >> 32
1184 dst.z = (src0.z \times src1.z) >> 32
1186 dst.w = (src0.w \times src1.w) >> 32
1189 .. opcode:: UMUL_HI - Unsigned Integer Multiply High Bits
1191 The high 32bits of the multiplication of 2 unsigned integers are returned.
1195 dst.x = (src0.x \times src1.x) >> 32
1197 dst.y = (src0.y \times src1.y) >> 32
1199 dst.z = (src0.z \times src1.z) >> 32
1201 dst.w = (src0.w \times src1.w) >> 32
1204 .. opcode:: IDIV - Signed Integer Division
1206 TBD: behavior for division by zero.
1210 dst.x = \frac{src0.x}{src1.x}
1212 dst.y = \frac{src0.y}{src1.y}
1214 dst.z = \frac{src0.z}{src1.z}
1216 dst.w = \frac{src0.w}{src1.w}
1219 .. opcode:: UDIV - Unsigned Integer Division
1221 For division by zero, 0xffffffff is returned.
1225 dst.x = \frac{src0.x}{src1.x}
1227 dst.y = \frac{src0.y}{src1.y}
1229 dst.z = \frac{src0.z}{src1.z}
1231 dst.w = \frac{src0.w}{src1.w}
1234 .. opcode:: UMOD - Unsigned Integer Remainder
1236 If second arg is zero, 0xffffffff is returned.
1240 dst.x = src0.x \bmod src1.x
1242 dst.y = src0.y \bmod src1.y
1244 dst.z = src0.z \bmod src1.z
1246 dst.w = src0.w \bmod src1.w
1249 .. opcode:: NOT - Bitwise Not
1262 .. opcode:: AND - Bitwise And
1266 dst.x = src0.x \& src1.x
1268 dst.y = src0.y \& src1.y
1270 dst.z = src0.z \& src1.z
1272 dst.w = src0.w \& src1.w
1275 .. opcode:: OR - Bitwise Or
1279 dst.x = src0.x | src1.x
1281 dst.y = src0.y | src1.y
1283 dst.z = src0.z | src1.z
1285 dst.w = src0.w | src1.w
1288 .. opcode:: XOR - Bitwise Xor
1292 dst.x = src0.x \oplus src1.x
1294 dst.y = src0.y \oplus src1.y
1296 dst.z = src0.z \oplus src1.z
1298 dst.w = src0.w \oplus src1.w
1301 .. opcode:: IMAX - Maximum of Signed Integers
1305 dst.x = max(src0.x, src1.x)
1307 dst.y = max(src0.y, src1.y)
1309 dst.z = max(src0.z, src1.z)
1311 dst.w = max(src0.w, src1.w)
1314 .. opcode:: UMAX - Maximum of Unsigned Integers
1318 dst.x = max(src0.x, src1.x)
1320 dst.y = max(src0.y, src1.y)
1322 dst.z = max(src0.z, src1.z)
1324 dst.w = max(src0.w, src1.w)
1327 .. opcode:: IMIN - Minimum of Signed Integers
1331 dst.x = min(src0.x, src1.x)
1333 dst.y = min(src0.y, src1.y)
1335 dst.z = min(src0.z, src1.z)
1337 dst.w = min(src0.w, src1.w)
1340 .. opcode:: UMIN - Minimum of Unsigned Integers
1344 dst.x = min(src0.x, src1.x)
1346 dst.y = min(src0.y, src1.y)
1348 dst.z = min(src0.z, src1.z)
1350 dst.w = min(src0.w, src1.w)
1353 .. opcode:: SHL - Shift Left
1355 The shift count is masked with 0x1f before the shift is applied.
1359 dst.x = src0.x << (0x1f \& src1.x)
1361 dst.y = src0.y << (0x1f \& src1.y)
1363 dst.z = src0.z << (0x1f \& src1.z)
1365 dst.w = src0.w << (0x1f \& src1.w)
1368 .. opcode:: ISHR - Arithmetic Shift Right (of Signed Integer)
1370 The shift count is masked with 0x1f before the shift is applied.
1374 dst.x = src0.x >> (0x1f \& src1.x)
1376 dst.y = src0.y >> (0x1f \& src1.y)
1378 dst.z = src0.z >> (0x1f \& src1.z)
1380 dst.w = src0.w >> (0x1f \& src1.w)
1383 .. opcode:: USHR - Logical Shift Right
1385 The shift count is masked with 0x1f before the shift is applied.
1389 dst.x = src0.x >> (unsigned) (0x1f \& src1.x)
1391 dst.y = src0.y >> (unsigned) (0x1f \& src1.y)
1393 dst.z = src0.z >> (unsigned) (0x1f \& src1.z)
1395 dst.w = src0.w >> (unsigned) (0x1f \& src1.w)
1398 .. opcode:: UCMP - Integer Conditional Move
1402 dst.x = src0.x ? src1.x : src2.x
1404 dst.y = src0.y ? src1.y : src2.y
1406 dst.z = src0.z ? src1.z : src2.z
1408 dst.w = src0.w ? src1.w : src2.w
1412 .. opcode:: ISSG - Integer Set Sign
1416 dst.x = (src0.x < 0) ? -1 : (src0.x > 0) ? 1 : 0
1418 dst.y = (src0.y < 0) ? -1 : (src0.y > 0) ? 1 : 0
1420 dst.z = (src0.z < 0) ? -1 : (src0.z > 0) ? 1 : 0
1422 dst.w = (src0.w < 0) ? -1 : (src0.w > 0) ? 1 : 0
1426 .. opcode:: FSLT - Float Set On Less Than (ordered)
1428 Same comparison as SLT but returns integer instead of 1.0/0.0 float
1432 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1434 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1436 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1438 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1441 .. opcode:: ISLT - Signed Integer Set On Less Than
1445 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1447 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1449 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1451 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1454 .. opcode:: USLT - Unsigned Integer Set On Less Than
1458 dst.x = (src0.x < src1.x) ? \sim 0 : 0
1460 dst.y = (src0.y < src1.y) ? \sim 0 : 0
1462 dst.z = (src0.z < src1.z) ? \sim 0 : 0
1464 dst.w = (src0.w < src1.w) ? \sim 0 : 0
1467 .. opcode:: FSGE - Float Set On Greater Equal Than (ordered)
1469 Same comparison as SGE but returns integer instead of 1.0/0.0 float
1473 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1475 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1477 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1479 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1482 .. opcode:: ISGE - Signed Integer Set On Greater Equal Than
1486 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1488 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1490 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1492 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1495 .. opcode:: USGE - Unsigned Integer Set On Greater Equal Than
1499 dst.x = (src0.x >= src1.x) ? \sim 0 : 0
1501 dst.y = (src0.y >= src1.y) ? \sim 0 : 0
1503 dst.z = (src0.z >= src1.z) ? \sim 0 : 0
1505 dst.w = (src0.w >= src1.w) ? \sim 0 : 0
1508 .. opcode:: FSEQ - Float Set On Equal (ordered)
1510 Same comparison as SEQ but returns integer instead of 1.0/0.0 float
1514 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1516 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1518 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1520 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1523 .. opcode:: USEQ - Integer Set On Equal
1527 dst.x = (src0.x == src1.x) ? \sim 0 : 0
1529 dst.y = (src0.y == src1.y) ? \sim 0 : 0
1531 dst.z = (src0.z == src1.z) ? \sim 0 : 0
1533 dst.w = (src0.w == src1.w) ? \sim 0 : 0
1536 .. opcode:: FSNE - Float Set On Not Equal (unordered)
1538 Same comparison as SNE but returns integer instead of 1.0/0.0 float
1542 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1544 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1546 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1548 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1551 .. opcode:: USNE - Integer Set On Not Equal
1555 dst.x = (src0.x != src1.x) ? \sim 0 : 0
1557 dst.y = (src0.y != src1.y) ? \sim 0 : 0
1559 dst.z = (src0.z != src1.z) ? \sim 0 : 0
1561 dst.w = (src0.w != src1.w) ? \sim 0 : 0
1564 .. opcode:: INEG - Integer Negate
1579 .. opcode:: IABS - Integer Absolute Value
1593 These opcodes are used for bit-level manipulation of integers.
1595 .. opcode:: IBFE - Signed Bitfield Extract
1597 Like GLSL bitfieldExtract. Extracts a set of bits from the input, and
1598 sign-extends them if the high bit of the extracted window is set.
1602 def ibfe(value, offset, bits):
1603 if offset < 0 or bits < 0 or offset + bits > 32:
1605 if bits == 0: return 0
1606 # Note: >> sign-extends
1607 return (value << (32 - offset - bits)) >> (32 - bits)
1609 .. opcode:: UBFE - Unsigned Bitfield Extract
1611 Like GLSL bitfieldExtract. Extracts a set of bits from the input, without
1616 def ubfe(value, offset, bits):
1617 if offset < 0 or bits < 0 or offset + bits > 32:
1619 if bits == 0: return 0
1620 # Note: >> does not sign-extend
1621 return (value << (32 - offset - bits)) >> (32 - bits)
1623 .. opcode:: BFI - Bitfield Insert
1625 Like GLSL bitfieldInsert. Replaces a bit region of 'base' with the low bits
1630 def bfi(base, insert, offset, bits):
1631 if offset < 0 or bits < 0 or offset + bits > 32:
1633 # << defined such that mask == ~0 when bits == 32, offset == 0
1634 mask = ((1 << bits) - 1) << offset
1635 return ((insert << offset) & mask) | (base & ~mask)
1637 .. opcode:: BREV - Bitfield Reverse
1639 See SM5 instruction BFREV. Reverses the bits of the argument.
1641 .. opcode:: POPC - Population Count
1643 See SM5 instruction COUNTBITS. Counts the number of set bits in the argument.
1645 .. opcode:: LSB - Index of lowest set bit
1647 See SM5 instruction FIRSTBIT_LO. Computes the 0-based index of the first set
1648 bit of the argument. Returns -1 if none are set.
1650 .. opcode:: IMSB - Index of highest non-sign bit
1652 See SM5 instruction FIRSTBIT_SHI. Computes the 0-based index of the highest
1653 non-sign bit of the argument (i.e. highest 0 bit for negative numbers,
1654 highest 1 bit for positive numbers). Returns -1 if all bits are the same
1655 (i.e. for inputs 0 and -1).
1657 .. opcode:: UMSB - Index of highest set bit
1659 See SM5 instruction FIRSTBIT_HI. Computes the 0-based index of the highest
1660 set bit of the argument. Returns -1 if none are set.
1663 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1665 These opcodes are only supported in geometry shaders; they have no meaning
1666 in any other type of shader.
1668 .. opcode:: EMIT - Emit
1670 Generate a new vertex for the current primitive into the specified vertex
1671 stream using the values in the output registers.
1674 .. opcode:: ENDPRIM - End Primitive
1676 Complete the current primitive in the specified vertex stream (consisting of
1677 the emitted vertices), and start a new one.
1683 These opcodes are part of :term:`GLSL`'s opcode set. Support for these
1684 opcodes is determined by a special capability bit, ``GLSL``.
1685 Some require glsl version 1.30 (UIF/BREAKC/SWITCH/CASE/DEFAULT/ENDSWITCH).
1687 .. opcode:: CAL - Subroutine Call
1693 .. opcode:: RET - Subroutine Call Return
1698 .. opcode:: CONT - Continue
1700 Unconditionally moves the point of execution to the instruction after the
1701 last bgnloop. The instruction must appear within a bgnloop/endloop.
1705 Support for CONT is determined by a special capability bit,
1706 ``TGSI_CONT_SUPPORTED``. See :ref:`Screen` for more information.
1709 .. opcode:: BGNLOOP - Begin a Loop
1711 Start a loop. Must have a matching endloop.
1714 .. opcode:: BGNSUB - Begin Subroutine
1716 Starts definition of a subroutine. Must have a matching endsub.
1719 .. opcode:: ENDLOOP - End a Loop
1721 End a loop started with bgnloop.
1724 .. opcode:: ENDSUB - End Subroutine
1726 Ends definition of a subroutine.
1729 .. opcode:: NOP - No Operation
1734 .. opcode:: BRK - Break
1736 Unconditionally moves the point of execution to the instruction after the
1737 next endloop or endswitch. The instruction must appear within a loop/endloop
1738 or switch/endswitch.
1741 .. opcode:: BREAKC - Break Conditional
1743 Conditionally moves the point of execution to the instruction after the
1744 next endloop or endswitch. The instruction must appear within a loop/endloop
1745 or switch/endswitch.
1746 Condition evaluates to true if src0.x != 0 where src0.x is interpreted
1747 as an integer register.
1751 Considered for removal as it's quite inconsistent wrt other opcodes
1752 (could emulate with UIF/BRK/ENDIF).
1755 .. opcode:: IF - Float If
1757 Start an IF ... ELSE .. ENDIF block. Condition evaluates to true if
1761 where src0.x is interpreted as a floating point register.
1764 .. opcode:: UIF - Bitwise If
1766 Start an UIF ... ELSE .. ENDIF block. Condition evaluates to true if
1770 where src0.x is interpreted as an integer register.
1773 .. opcode:: ELSE - Else
1775 Starts an else block, after an IF or UIF statement.
1778 .. opcode:: ENDIF - End If
1780 Ends an IF or UIF block.
1783 .. opcode:: SWITCH - Switch
1785 Starts a C-style switch expression. The switch consists of one or multiple
1786 CASE statements, and at most one DEFAULT statement. Execution of a statement
1787 ends when a BRK is hit, but just like in C falling through to other cases
1788 without a break is allowed. Similarly, DEFAULT label is allowed anywhere not
1789 just as last statement, and fallthrough is allowed into/from it.
1790 CASE src arguments are evaluated at bit level against the SWITCH src argument.
1796 (some instructions here)
1799 (some instructions here)
1802 (some instructions here)
1807 .. opcode:: CASE - Switch case
1809 This represents a switch case label. The src arg must be an integer immediate.
1812 .. opcode:: DEFAULT - Switch default
1814 This represents the default case in the switch, which is taken if no other
1818 .. opcode:: ENDSWITCH - End of switch
1820 Ends a switch expression.
1826 The interpolation instructions allow an input to be interpolated in a
1827 different way than its declaration. This corresponds to the GLSL 4.00
1828 interpolateAt* functions. The first argument of each of these must come from
1829 ``TGSI_FILE_INPUT``.
1831 .. opcode:: INTERP_CENTROID - Interpolate at the centroid
1833 Interpolates the varying specified by src0 at the centroid
1835 .. opcode:: INTERP_SAMPLE - Interpolate at the specified sample
1837 Interpolates the varying specified by src0 at the sample id specified by
1838 src1.x (interpreted as an integer)
1840 .. opcode:: INTERP_OFFSET - Interpolate at the specified offset
1842 Interpolates the varying specified by src0 at the offset src1.xy from the
1843 pixel center (interpreted as floats)
1851 The double-precision opcodes reinterpret four-component vectors into
1852 two-component vectors with doubled precision in each component.
1854 .. opcode:: DABS - Absolute
1862 .. opcode:: DADD - Add
1866 dst.xy = src0.xy + src1.xy
1868 dst.zw = src0.zw + src1.zw
1870 .. opcode:: DSEQ - Set on Equal
1874 dst.x = src0.xy == src1.xy ? \sim 0 : 0
1876 dst.z = src0.zw == src1.zw ? \sim 0 : 0
1878 .. opcode:: DSNE - Set on Equal
1882 dst.x = src0.xy != src1.xy ? \sim 0 : 0
1884 dst.z = src0.zw != src1.zw ? \sim 0 : 0
1886 .. opcode:: DSLT - Set on Less than
1890 dst.x = src0.xy < src1.xy ? \sim 0 : 0
1892 dst.z = src0.zw < src1.zw ? \sim 0 : 0
1894 .. opcode:: DSGE - Set on Greater equal
1898 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
1900 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
1902 .. opcode:: DFRAC - Fraction
1906 dst.xy = src.xy - \lfloor src.xy\rfloor
1908 dst.zw = src.zw - \lfloor src.zw\rfloor
1910 .. opcode:: DTRUNC - Truncate
1914 dst.xy = trunc(src.xy)
1916 dst.zw = trunc(src.zw)
1918 .. opcode:: DCEIL - Ceiling
1922 dst.xy = \lceil src.xy\rceil
1924 dst.zw = \lceil src.zw\rceil
1926 .. opcode:: DFLR - Floor
1930 dst.xy = \lfloor src.xy\rfloor
1932 dst.zw = \lfloor src.zw\rfloor
1934 .. opcode:: DROUND - Fraction
1938 dst.xy = round(src.xy)
1940 dst.zw = round(src.zw)
1942 .. opcode:: DSSG - Set Sign
1946 dst.xy = (src.xy > 0) ? 1.0 : (src.xy < 0) ? -1.0 : 0.0
1948 dst.zw = (src.zw > 0) ? 1.0 : (src.zw < 0) ? -1.0 : 0.0
1950 .. opcode:: DFRACEXP - Convert Number to Fractional and Integral Components
1952 Like the ``frexp()`` routine in many math libraries, this opcode stores the
1953 exponent of its source to ``dst0``, and the significand to ``dst1``, such that
1954 :math:`dst1 \times 2^{dst0} = src` .
1958 dst0.xy = exp(src.xy)
1960 dst1.xy = frac(src.xy)
1962 dst0.zw = exp(src.zw)
1964 dst1.zw = frac(src.zw)
1966 .. opcode:: DLDEXP - Multiply Number by Integral Power of 2
1968 This opcode is the inverse of :opcode:`DFRACEXP`. The second
1969 source is an integer.
1973 dst.xy = src0.xy \times 2^{src1.x}
1975 dst.zw = src0.zw \times 2^{src1.y}
1977 .. opcode:: DMIN - Minimum
1981 dst.xy = min(src0.xy, src1.xy)
1983 dst.zw = min(src0.zw, src1.zw)
1985 .. opcode:: DMAX - Maximum
1989 dst.xy = max(src0.xy, src1.xy)
1991 dst.zw = max(src0.zw, src1.zw)
1993 .. opcode:: DMUL - Multiply
1997 dst.xy = src0.xy \times src1.xy
1999 dst.zw = src0.zw \times src1.zw
2002 .. opcode:: DMAD - Multiply And Add
2006 dst.xy = src0.xy \times src1.xy + src2.xy
2008 dst.zw = src0.zw \times src1.zw + src2.zw
2011 .. opcode:: DFMA - Fused Multiply-Add
2013 Perform a * b + c with no intermediate rounding step.
2017 dst.xy = src0.xy \times src1.xy + src2.xy
2019 dst.zw = src0.zw \times src1.zw + src2.zw
2022 .. opcode:: DDIV - Divide
2026 dst.xy = \frac{src0.xy}{src1.xy}
2028 dst.zw = \frac{src0.zw}{src1.zw}
2031 .. opcode:: DRCP - Reciprocal
2035 dst.xy = \frac{1}{src.xy}
2037 dst.zw = \frac{1}{src.zw}
2039 .. opcode:: DSQRT - Square Root
2043 dst.xy = \sqrt{src.xy}
2045 dst.zw = \sqrt{src.zw}
2047 .. opcode:: DRSQ - Reciprocal Square Root
2051 dst.xy = \frac{1}{\sqrt{src.xy}}
2053 dst.zw = \frac{1}{\sqrt{src.zw}}
2055 .. opcode:: F2D - Float to Double
2059 dst.xy = double(src0.x)
2061 dst.zw = double(src0.y)
2063 .. opcode:: D2F - Double to Float
2067 dst.x = float(src0.xy)
2069 dst.y = float(src0.zw)
2071 .. opcode:: I2D - Int to Double
2075 dst.xy = double(src0.x)
2077 dst.zw = double(src0.y)
2079 .. opcode:: D2I - Double to Int
2083 dst.x = int(src0.xy)
2085 dst.y = int(src0.zw)
2087 .. opcode:: U2D - Unsigned Int to Double
2091 dst.xy = double(src0.x)
2093 dst.zw = double(src0.y)
2095 .. opcode:: D2U - Double to Unsigned Int
2099 dst.x = unsigned(src0.xy)
2101 dst.y = unsigned(src0.zw)
2106 The 64-bit integer opcodes reinterpret four-component vectors into
2107 two-component vectors with 64-bits in each component.
2109 .. opcode:: I64ABS - 64-bit Integer Absolute Value
2117 .. opcode:: I64NEG - 64-bit Integer Negate
2127 .. opcode:: I64SSG - 64-bit Integer Set Sign
2131 dst.xy = (src0.xy < 0) ? -1 : (src0.xy > 0) ? 1 : 0
2133 dst.zw = (src0.zw < 0) ? -1 : (src0.zw > 0) ? 1 : 0
2135 .. opcode:: U64ADD - 64-bit Integer Add
2139 dst.xy = src0.xy + src1.xy
2141 dst.zw = src0.zw + src1.zw
2143 .. opcode:: U64MUL - 64-bit Integer Multiply
2147 dst.xy = src0.xy * src1.xy
2149 dst.zw = src0.zw * src1.zw
2151 .. opcode:: U64SEQ - 64-bit Integer Set on Equal
2155 dst.x = src0.xy == src1.xy ? \sim 0 : 0
2157 dst.z = src0.zw == src1.zw ? \sim 0 : 0
2159 .. opcode:: U64SNE - 64-bit Integer Set on Not Equal
2163 dst.x = src0.xy != src1.xy ? \sim 0 : 0
2165 dst.z = src0.zw != src1.zw ? \sim 0 : 0
2167 .. opcode:: U64SLT - 64-bit Unsigned Integer Set on Less Than
2171 dst.x = src0.xy < src1.xy ? \sim 0 : 0
2173 dst.z = src0.zw < src1.zw ? \sim 0 : 0
2175 .. opcode:: U64SGE - 64-bit Unsigned Integer Set on Greater Equal
2179 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
2181 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
2183 .. opcode:: I64SLT - 64-bit Signed Integer Set on Less Than
2187 dst.x = src0.xy < src1.xy ? \sim 0 : 0
2189 dst.z = src0.zw < src1.zw ? \sim 0 : 0
2191 .. opcode:: I64SGE - 64-bit Signed Integer Set on Greater Equal
2195 dst.x = src0.xy >= src1.xy ? \sim 0 : 0
2197 dst.z = src0.zw >= src1.zw ? \sim 0 : 0
2199 .. opcode:: I64MIN - Minimum of 64-bit Signed Integers
2203 dst.xy = min(src0.xy, src1.xy)
2205 dst.zw = min(src0.zw, src1.zw)
2207 .. opcode:: U64MIN - Minimum of 64-bit Unsigned Integers
2211 dst.xy = min(src0.xy, src1.xy)
2213 dst.zw = min(src0.zw, src1.zw)
2215 .. opcode:: I64MAX - Maximum of 64-bit Signed Integers
2219 dst.xy = max(src0.xy, src1.xy)
2221 dst.zw = max(src0.zw, src1.zw)
2223 .. opcode:: U64MAX - Maximum of 64-bit Unsigned Integers
2227 dst.xy = max(src0.xy, src1.xy)
2229 dst.zw = max(src0.zw, src1.zw)
2231 .. opcode:: U64SHL - Shift Left 64-bit Unsigned Integer
2233 The shift count is masked with 0x3f before the shift is applied.
2237 dst.xy = src0.xy << (0x3f \& src1.x)
2239 dst.zw = src0.zw << (0x3f \& src1.y)
2241 .. opcode:: I64SHR - Arithmetic Shift Right (of 64-bit Signed Integer)
2243 The shift count is masked with 0x3f before the shift is applied.
2247 dst.xy = src0.xy >> (0x3f \& src1.x)
2249 dst.zw = src0.zw >> (0x3f \& src1.y)
2251 .. opcode:: U64SHR - Logical Shift Right (of 64-bit Unsigned Integer)
2253 The shift count is masked with 0x3f before the shift is applied.
2257 dst.xy = src0.xy >> (unsigned) (0x3f \& src1.x)
2259 dst.zw = src0.zw >> (unsigned) (0x3f \& src1.y)
2261 .. opcode:: I64DIV - 64-bit Signed Integer Division
2265 dst.xy = \frac{src0.xy}{src1.xy}
2267 dst.zw = \frac{src0.zw}{src1.zw}
2269 .. opcode:: U64DIV - 64-bit Unsigned Integer Division
2273 dst.xy = \frac{src0.xy}{src1.xy}
2275 dst.zw = \frac{src0.zw}{src1.zw}
2277 .. opcode:: U64MOD - 64-bit Unsigned Integer Remainder
2281 dst.xy = src0.xy \bmod src1.xy
2283 dst.zw = src0.zw \bmod src1.zw
2285 .. opcode:: I64MOD - 64-bit Signed Integer Remainder
2289 dst.xy = src0.xy \bmod src1.xy
2291 dst.zw = src0.zw \bmod src1.zw
2293 .. opcode:: F2U64 - Float to 64-bit Unsigned Int
2297 dst.xy = (uint64_t) src0.x
2299 dst.zw = (uint64_t) src0.y
2301 .. opcode:: F2I64 - Float to 64-bit Int
2305 dst.xy = (int64_t) src0.x
2307 dst.zw = (int64_t) src0.y
2309 .. opcode:: U2I64 - Unsigned Integer to 64-bit Integer
2311 This is a zero extension.
2315 dst.xy = (uint64_t) src0.x
2317 dst.zw = (uint64_t) src0.y
2319 .. opcode:: I2I64 - Signed Integer to 64-bit Integer
2321 This is a sign extension.
2325 dst.xy = (int64_t) src0.x
2327 dst.zw = (int64_t) src0.y
2329 .. opcode:: D2U64 - Double to 64-bit Unsigned Int
2333 dst.xy = (uint64_t) src0.xy
2335 dst.zw = (uint64_t) src0.zw
2337 .. opcode:: D2I64 - Double to 64-bit Int
2341 dst.xy = (int64_t) src0.xy
2343 dst.zw = (int64_t) src0.zw
2345 .. opcode:: U642F - 64-bit unsigned integer to float
2349 dst.x = (float) src0.xy
2351 dst.y = (float) src0.zw
2353 .. opcode:: I642F - 64-bit Int to Float
2357 dst.x = (float) src0.xy
2359 dst.y = (float) src0.zw
2361 .. opcode:: U642D - 64-bit unsigned integer to double
2365 dst.xy = (double) src0.xy
2367 dst.zw = (double) src0.zw
2369 .. opcode:: I642D - 64-bit Int to double
2373 dst.xy = (double) src0.xy
2375 dst.zw = (double) src0.zw
2377 .. _samplingopcodes:
2379 Resource Sampling Opcodes
2380 ^^^^^^^^^^^^^^^^^^^^^^^^^
2382 Those opcodes follow very closely semantics of the respective Direct3D
2383 instructions. If in doubt double check Direct3D documentation.
2384 Note that the swizzle on SVIEW (src1) determines texel swizzling
2389 Using provided address, sample data from the specified texture using the
2390 filtering mode identified by the given sampler. The source data may come from
2391 any resource type other than buffers.
2393 Syntax: ``SAMPLE dst, address, sampler_view, sampler``
2395 Example: ``SAMPLE TEMP[0], TEMP[1], SVIEW[0], SAMP[0]``
2397 .. opcode:: SAMPLE_I
2399 Simplified alternative to the SAMPLE instruction. Using the provided
2400 integer address, SAMPLE_I fetches data from the specified sampler view
2401 without any filtering. The source data may come from any resource type
2404 Syntax: ``SAMPLE_I dst, address, sampler_view``
2406 Example: ``SAMPLE_I TEMP[0], TEMP[1], SVIEW[0]``
2408 The 'address' is specified as unsigned integers. If the 'address' is out of
2409 range [0...(# texels - 1)] the result of the fetch is always 0 in all
2410 components. As such the instruction doesn't honor address wrap modes, in
2411 cases where that behavior is desirable 'SAMPLE' instruction should be used.
2412 address.w always provides an unsigned integer mipmap level. If the value is
2413 out of the range then the instruction always returns 0 in all components.
2414 address.yz are ignored for buffers and 1d textures. address.z is ignored
2415 for 1d texture arrays and 2d textures.
2417 For 1D texture arrays address.y provides the array index (also as unsigned
2418 integer). If the value is out of the range of available array indices
2419 [0... (array size - 1)] then the opcode always returns 0 in all components.
2420 For 2D texture arrays address.z provides the array index, otherwise it
2421 exhibits the same behavior as in the case for 1D texture arrays. The exact
2422 semantics of the source address are presented in the table below:
2424 +---------------------------+----+-----+-----+---------+
2425 | resource type | X | Y | Z | W |
2426 +===========================+====+=====+=====+=========+
2427 | ``PIPE_BUFFER`` | x | | | ignored |
2428 +---------------------------+----+-----+-----+---------+
2429 | ``PIPE_TEXTURE_1D`` | x | | | mpl |
2430 +---------------------------+----+-----+-----+---------+
2431 | ``PIPE_TEXTURE_2D`` | x | y | | mpl |
2432 +---------------------------+----+-----+-----+---------+
2433 | ``PIPE_TEXTURE_3D`` | x | y | z | mpl |
2434 +---------------------------+----+-----+-----+---------+
2435 | ``PIPE_TEXTURE_RECT`` | x | y | | mpl |
2436 +---------------------------+----+-----+-----+---------+
2437 | ``PIPE_TEXTURE_CUBE`` | not allowed as source |
2438 +---------------------------+----+-----+-----+---------+
2439 | ``PIPE_TEXTURE_1D_ARRAY`` | x | idx | | mpl |
2440 +---------------------------+----+-----+-----+---------+
2441 | ``PIPE_TEXTURE_2D_ARRAY`` | x | y | idx | mpl |
2442 +---------------------------+----+-----+-----+---------+
2444 Where 'mpl' is a mipmap level and 'idx' is the array index.
2446 .. opcode:: SAMPLE_I_MS
2448 Just like SAMPLE_I but allows fetch data from multi-sampled surfaces.
2450 Syntax: ``SAMPLE_I_MS dst, address, sampler_view, sample``
2452 .. opcode:: SAMPLE_B
2454 Just like the SAMPLE instruction with the exception that an additional bias
2455 is applied to the level of detail computed as part of the instruction
2458 Syntax: ``SAMPLE_B dst, address, sampler_view, sampler, lod_bias``
2460 Example: ``SAMPLE_B TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2462 .. opcode:: SAMPLE_C
2464 Similar to the SAMPLE instruction but it performs a comparison filter. The
2465 operands to SAMPLE_C are identical to SAMPLE, except that there is an
2466 additional float32 operand, reference value, which must be a register with
2467 single-component, or a scalar literal. SAMPLE_C makes the hardware use the
2468 current samplers compare_func (in pipe_sampler_state) to compare reference
2469 value against the red component value for the surce resource at each texel
2470 that the currently configured texture filter covers based on the provided
2473 Syntax: ``SAMPLE_C dst, address, sampler_view.r, sampler, ref_value``
2475 Example: ``SAMPLE_C TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2477 .. opcode:: SAMPLE_C_LZ
2479 Same as SAMPLE_C, but LOD is 0 and derivatives are ignored. The LZ stands
2482 Syntax: ``SAMPLE_C_LZ dst, address, sampler_view.r, sampler, ref_value``
2484 Example: ``SAMPLE_C_LZ TEMP[0], TEMP[1], SVIEW[0].r, SAMP[0], TEMP[2].x``
2487 .. opcode:: SAMPLE_D
2489 SAMPLE_D is identical to the SAMPLE opcode except that the derivatives for
2490 the source address in the x direction and the y direction are provided by
2493 Syntax: ``SAMPLE_D dst, address, sampler_view, sampler, der_x, der_y``
2495 Example: ``SAMPLE_D TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2], TEMP[3]``
2497 .. opcode:: SAMPLE_L
2499 SAMPLE_L is identical to the SAMPLE opcode except that the LOD is provided
2500 directly as a scalar value, representing no anisotropy.
2502 Syntax: ``SAMPLE_L dst, address, sampler_view, sampler, explicit_lod``
2504 Example: ``SAMPLE_L TEMP[0], TEMP[1], SVIEW[0], SAMP[0], TEMP[2].x``
2508 Gathers the four texels to be used in a bi-linear filtering operation and
2509 packs them into a single register. Only works with 2D, 2D array, cubemaps,
2510 and cubemaps arrays. For 2D textures, only the addressing modes of the
2511 sampler and the top level of any mip pyramid are used. Set W to zero. It
2512 behaves like the SAMPLE instruction, but a filtered sample is not
2513 generated. The four samples that contribute to filtering are placed into
2514 xyzw in counter-clockwise order, starting with the (u,v) texture coordinate
2515 delta at the following locations (-, +), (+, +), (+, -), (-, -), where the
2516 magnitude of the deltas are half a texel.
2519 .. opcode:: SVIEWINFO
2521 Query the dimensions of a given sampler view. dst receives width, height,
2522 depth or array size and number of mipmap levels as int4. The dst can have a
2523 writemask which will specify what info is the caller interested in.
2525 Syntax: ``SVIEWINFO dst, src_mip_level, sampler_view``
2527 Example: ``SVIEWINFO TEMP[0], TEMP[1].x, SVIEW[0]``
2529 src_mip_level is an unsigned integer scalar. If it's out of range then
2530 returns 0 for width, height and depth/array size but the total number of
2531 mipmap is still returned correctly for the given sampler view. The returned
2532 width, height and depth values are for the mipmap level selected by the
2533 src_mip_level and are in the number of texels. For 1d texture array width
2534 is in dst.x, array size is in dst.y and dst.z is 0. The number of mipmaps is
2535 still in dst.w. In contrast to d3d10 resinfo, there's no way in the tgsi
2536 instruction encoding to specify the return type (float/rcpfloat/uint), hence
2537 always using uint. Also, unlike the SAMPLE instructions, the swizzle on src1
2538 resinfo allowing swizzling dst values is ignored (due to the interaction
2539 with rcpfloat modifier which requires some swizzle handling in the state
2542 .. opcode:: SAMPLE_POS
2544 Query the position of a sample in the given resource or render target
2545 when per-sample fragment shading is in effect.
2547 Syntax: ``SAMPLE_POS dst, source, sample_index``
2549 dst receives float4 (x, y, undef, undef) indicated where the sample is
2550 located. Sample locations are in the range [0, 1] where 0.5 is the center
2553 source is either a sampler view (to indicate a shader resource) or temp
2554 register (to indicate the render target). The source register may have
2555 an optional swizzle to apply to the returned result
2557 sample_index is an integer scalar indicating which sample position is to
2560 If per-sample shading is not in effect or the source resource or render
2561 target is not multisampled, the result is (0.5, 0.5, undef, undef).
2563 NOTE: no driver has implemented this opcode yet (and no state tracker
2564 emits it). This information is subject to change.
2566 .. opcode:: SAMPLE_INFO
2568 Query the number of samples in a multisampled resource or render target.
2570 Syntax: ``SAMPLE_INFO dst, source``
2572 dst receives int4 (n, 0, 0, 0) where n is the number of samples in a
2573 resource or the render target.
2575 source is either a sampler view (to indicate a shader resource) or temp
2576 register (to indicate the render target). The source register may have
2577 an optional swizzle to apply to the returned result
2579 If per-sample shading is not in effect or the source resource or render
2580 target is not multisampled, the result is (1, 0, 0, 0).
2582 NOTE: no driver has implemented this opcode yet (and no state tracker
2583 emits it). This information is subject to change.
2585 .. _resourceopcodes:
2587 Resource Access Opcodes
2588 ^^^^^^^^^^^^^^^^^^^^^^^
2590 For these opcodes, the resource can be a BUFFER, IMAGE, or MEMORY.
2592 .. opcode:: LOAD - Fetch data from a shader buffer or image
2594 Syntax: ``LOAD dst, resource, address``
2596 Example: ``LOAD TEMP[0], BUFFER[0], TEMP[1]``
2598 Using the provided integer address, LOAD fetches data
2599 from the specified buffer or texture without any
2602 The 'address' is specified as a vector of unsigned
2603 integers. If the 'address' is out of range the result
2606 Only the first mipmap level of a resource can be read
2607 from using this instruction.
2609 For 1D or 2D texture arrays, the array index is
2610 provided as an unsigned integer in address.y or
2611 address.z, respectively. address.yz are ignored for
2612 buffers and 1D textures. address.z is ignored for 1D
2613 texture arrays and 2D textures. address.w is always
2616 A swizzle suffix may be added to the resource argument
2617 this will cause the resource data to be swizzled accordingly.
2619 .. opcode:: STORE - Write data to a shader resource
2621 Syntax: ``STORE resource, address, src``
2623 Example: ``STORE BUFFER[0], TEMP[0], TEMP[1]``
2625 Using the provided integer address, STORE writes data
2626 to the specified buffer or texture.
2628 The 'address' is specified as a vector of unsigned
2629 integers. If the 'address' is out of range the result
2632 Only the first mipmap level of a resource can be
2633 written to using this instruction.
2635 For 1D or 2D texture arrays, the array index is
2636 provided as an unsigned integer in address.y or
2637 address.z, respectively. address.yz are ignored for
2638 buffers and 1D textures. address.z is ignored for 1D
2639 texture arrays and 2D textures. address.w is always
2642 .. opcode:: RESQ - Query information about a resource
2644 Syntax: ``RESQ dst, resource``
2646 Example: ``RESQ TEMP[0], BUFFER[0]``
2648 Returns information about the buffer or image resource. For buffer
2649 resources, the size (in bytes) is returned in the x component. For
2650 image resources, .xyz will contain the width/height/layers of the
2651 image, while .w will contain the number of samples for multi-sampled
2654 .. opcode:: FBFETCH - Load data from framebuffer
2656 Syntax: ``FBFETCH dst, output``
2658 Example: ``FBFETCH TEMP[0], OUT[0]``
2660 This is only valid on ``COLOR`` semantic outputs. Returns the color
2661 of the current position in the framebuffer from before this fragment
2662 shader invocation. May return the same value from multiple calls for
2663 a particular output within a single invocation. Note that result may
2664 be undefined if a fragment is drawn multiple times without a blend
2668 .. _threadsyncopcodes:
2670 Inter-thread synchronization opcodes
2671 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2673 These opcodes are intended for communication between threads running
2674 within the same compute grid. For now they're only valid in compute
2677 .. opcode:: MFENCE - Memory fence
2679 Syntax: ``MFENCE resource``
2681 Example: ``MFENCE RES[0]``
2683 This opcode forces strong ordering between any memory access
2684 operations that affect the specified resource. This means that
2685 previous loads and stores (and only those) will be performed and
2686 visible to other threads before the program execution continues.
2689 .. opcode:: LFENCE - Load memory fence
2691 Syntax: ``LFENCE resource``
2693 Example: ``LFENCE RES[0]``
2695 Similar to MFENCE, but it only affects the ordering of memory loads.
2698 .. opcode:: SFENCE - Store memory fence
2700 Syntax: ``SFENCE resource``
2702 Example: ``SFENCE RES[0]``
2704 Similar to MFENCE, but it only affects the ordering of memory stores.
2707 .. opcode:: BARRIER - Thread group barrier
2711 This opcode suspends the execution of the current thread until all
2712 the remaining threads in the working group reach the same point of
2713 the program. Results are unspecified if any of the remaining
2714 threads terminates or never reaches an executed BARRIER instruction.
2716 .. opcode:: MEMBAR - Memory barrier
2720 This opcode waits for the completion of all memory accesses based on
2721 the type passed in. The type is an immediate bitfield with the following
2724 Bit 0: Shader storage buffers
2725 Bit 1: Atomic buffers
2727 Bit 3: Shared memory
2730 These may be passed in in any combination. An implementation is free to not
2731 distinguish between these as it sees fit. However these map to all the
2732 possibilities made available by GLSL.
2739 These opcodes provide atomic variants of some common arithmetic and
2740 logical operations. In this context atomicity means that another
2741 concurrent memory access operation that affects the same memory
2742 location is guaranteed to be performed strictly before or after the
2743 entire execution of the atomic operation. The resource may be a BUFFER,
2744 IMAGE, or MEMORY. In the case of an image, the offset works the same as for
2745 ``LOAD`` and ``STORE``, specified above. These atomic operations may
2746 only be used with 32-bit integer image formats.
2748 .. opcode:: ATOMUADD - Atomic integer addition
2750 Syntax: ``ATOMUADD dst, resource, offset, src``
2752 Example: ``ATOMUADD TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2754 The following operation is performed atomically:
2758 dst_x = resource[offset]
2760 resource[offset] = dst_x + src_x
2763 .. opcode:: ATOMXCHG - Atomic exchange
2765 Syntax: ``ATOMXCHG dst, resource, offset, src``
2767 Example: ``ATOMXCHG TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2769 The following operation is performed atomically:
2773 dst_x = resource[offset]
2775 resource[offset] = src_x
2778 .. opcode:: ATOMCAS - Atomic compare-and-exchange
2780 Syntax: ``ATOMCAS dst, resource, offset, cmp, src``
2782 Example: ``ATOMCAS TEMP[0], BUFFER[0], TEMP[1], TEMP[2], TEMP[3]``
2784 The following operation is performed atomically:
2788 dst_x = resource[offset]
2790 resource[offset] = (dst_x == cmp_x ? src_x : dst_x)
2793 .. opcode:: ATOMAND - Atomic bitwise And
2795 Syntax: ``ATOMAND dst, resource, offset, src``
2797 Example: ``ATOMAND TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2799 The following operation is performed atomically:
2803 dst_x = resource[offset]
2805 resource[offset] = dst_x \& src_x
2808 .. opcode:: ATOMOR - Atomic bitwise Or
2810 Syntax: ``ATOMOR dst, resource, offset, src``
2812 Example: ``ATOMOR TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2814 The following operation is performed atomically:
2818 dst_x = resource[offset]
2820 resource[offset] = dst_x | src_x
2823 .. opcode:: ATOMXOR - Atomic bitwise Xor
2825 Syntax: ``ATOMXOR dst, resource, offset, src``
2827 Example: ``ATOMXOR TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2829 The following operation is performed atomically:
2833 dst_x = resource[offset]
2835 resource[offset] = dst_x \oplus src_x
2838 .. opcode:: ATOMUMIN - Atomic unsigned minimum
2840 Syntax: ``ATOMUMIN dst, resource, offset, src``
2842 Example: ``ATOMUMIN TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2844 The following operation is performed atomically:
2848 dst_x = resource[offset]
2850 resource[offset] = (dst_x < src_x ? dst_x : src_x)
2853 .. opcode:: ATOMUMAX - Atomic unsigned maximum
2855 Syntax: ``ATOMUMAX dst, resource, offset, src``
2857 Example: ``ATOMUMAX TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2859 The following operation is performed atomically:
2863 dst_x = resource[offset]
2865 resource[offset] = (dst_x > src_x ? dst_x : src_x)
2868 .. opcode:: ATOMIMIN - Atomic signed minimum
2870 Syntax: ``ATOMIMIN dst, resource, offset, src``
2872 Example: ``ATOMIMIN TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2874 The following operation is performed atomically:
2878 dst_x = resource[offset]
2880 resource[offset] = (dst_x < src_x ? dst_x : src_x)
2883 .. opcode:: ATOMIMAX - Atomic signed maximum
2885 Syntax: ``ATOMIMAX dst, resource, offset, src``
2887 Example: ``ATOMIMAX TEMP[0], BUFFER[0], TEMP[1], TEMP[2]``
2889 The following operation is performed atomically:
2893 dst_x = resource[offset]
2895 resource[offset] = (dst_x > src_x ? dst_x : src_x)
2898 .. _interlaneopcodes:
2903 These opcodes reduce the given value across the shader invocations
2904 running in the current SIMD group. Every thread in the subgroup will receive
2905 the same result. The BALLOT operations accept a single-channel argument that
2906 is treated as a boolean and produce a 64-bit value.
2908 .. opcode:: VOTE_ANY - Value is set in any of the active invocations
2910 Syntax: ``VOTE_ANY dst, value``
2912 Example: ``VOTE_ANY TEMP[0].x, TEMP[1].x``
2915 .. opcode:: VOTE_ALL - Value is set in all of the active invocations
2917 Syntax: ``VOTE_ALL dst, value``
2919 Example: ``VOTE_ALL TEMP[0].x, TEMP[1].x``
2922 .. opcode:: VOTE_EQ - Value is the same in all of the active invocations
2924 Syntax: ``VOTE_EQ dst, value``
2926 Example: ``VOTE_EQ TEMP[0].x, TEMP[1].x``
2929 .. opcode:: BALLOT - Lanemask of whether the value is set in each active
2932 Syntax: ``BALLOT dst, value``
2934 Example: ``BALLOT TEMP[0].xy, TEMP[1].x``
2936 When the argument is a constant true, this produces a bitmask of active
2937 invocations. In fragment shaders, this can include helper invocations
2938 (invocations whose outputs and writes to memory are discarded, but which
2939 are used to compute derivatives).
2942 .. opcode:: READ_FIRST - Broadcast the value from the first active
2943 invocation to all active lanes
2945 Syntax: ``READ_FIRST dst, value``
2947 Example: ``READ_FIRST TEMP[0], TEMP[1]``
2950 .. opcode:: READ_INVOC - Retrieve the value from the given invocation
2951 (need not be uniform)
2953 Syntax: ``READ_INVOC dst, value, invocation``
2955 Example: ``READ_INVOC TEMP[0].xy, TEMP[1].xy, TEMP[2].x``
2957 invocation.x controls the invocation number to read from for all channels.
2958 The invocation number must be the same across all active invocations in a
2959 sub-group; otherwise, the results are undefined.
2962 Explanation of symbols used
2963 ------------------------------
2970 :math:`|x|` Absolute value of `x`.
2972 :math:`\lceil x \rceil` Ceiling of `x`.
2974 clamp(x,y,z) Clamp x between y and z.
2975 (x < y) ? y : (x > z) ? z : x
2977 :math:`\lfloor x\rfloor` Floor of `x`.
2979 :math:`\log_2{x}` Logarithm of `x`, base 2.
2981 max(x,y) Maximum of x and y.
2984 min(x,y) Minimum of x and y.
2987 partialx(x) Derivative of x relative to fragment's X.
2989 partialy(x) Derivative of x relative to fragment's Y.
2991 pop() Pop from stack.
2993 :math:`x^y` `x` to the power `y`.
2995 push(x) Push x on stack.
2999 trunc(x) Truncate x, i.e. drop the fraction bits.
3006 discard Discard fragment.
3010 target Label of target instruction.
3021 Declares a register that is will be referenced as an operand in Instruction
3024 File field contains register file that is being declared and is one
3027 UsageMask field specifies which of the register components can be accessed
3028 and is one of TGSI_WRITEMASK.
3030 The Local flag specifies that a given value isn't intended for
3031 subroutine parameter passing and, as a result, the implementation
3032 isn't required to give any guarantees of it being preserved across
3033 subroutine boundaries. As it's merely a compiler hint, the
3034 implementation is free to ignore it.
3036 If Dimension flag is set to 1, a Declaration Dimension token follows.
3038 If Semantic flag is set to 1, a Declaration Semantic token follows.
3040 If Interpolate flag is set to 1, a Declaration Interpolate token follows.
3042 If file is TGSI_FILE_RESOURCE, a Declaration Resource token follows.
3044 If Array flag is set to 1, a Declaration Array token follows.
3047 ^^^^^^^^^^^^^^^^^^^^^^^^
3049 Declarations can optional have an ArrayID attribute which can be referred by
3050 indirect addressing operands. An ArrayID of zero is reserved and treated as
3051 if no ArrayID is specified.
3053 If an indirect addressing operand refers to a specific declaration by using
3054 an ArrayID only the registers in this declaration are guaranteed to be
3055 accessed, accessing any register outside this declaration results in undefined
3056 behavior. Note that for compatibility the effective index is zero-based and
3057 not relative to the specified declaration
3059 If no ArrayID is specified with an indirect addressing operand the whole
3060 register file might be accessed by this operand. This is strongly discouraged
3061 and will prevent packing of scalar/vec2 arrays and effective alias analysis.
3062 This is only legal for TEMP and CONST register files.
3064 Declaration Semantic
3065 ^^^^^^^^^^^^^^^^^^^^^^^^
3067 Vertex and fragment shader input and output registers may be labeled
3068 with semantic information consisting of a name and index.
3070 Follows Declaration token if Semantic bit is set.
3072 Since its purpose is to link a shader with other stages of the pipeline,
3073 it is valid to follow only those Declaration tokens that declare a register
3074 either in INPUT or OUTPUT file.
3076 SemanticName field contains the semantic name of the register being declared.
3077 There is no default value.
3079 SemanticIndex is an optional subscript that can be used to distinguish
3080 different register declarations with the same semantic name. The default value
3083 The meanings of the individual semantic names are explained in the following
3086 TGSI_SEMANTIC_POSITION
3087 """"""""""""""""""""""
3089 For vertex shaders, TGSI_SEMANTIC_POSITION indicates the vertex shader
3090 output register which contains the homogeneous vertex position in the clip
3091 space coordinate system. After clipping, the X, Y and Z components of the
3092 vertex will be divided by the W value to get normalized device coordinates.
3094 For fragment shaders, TGSI_SEMANTIC_POSITION is used to indicate that
3095 fragment shader input (or system value, depending on which one is
3096 supported by the driver) contains the fragment's window position. The X
3097 component starts at zero and always increases from left to right.
3098 The Y component starts at zero and always increases but Y=0 may either
3099 indicate the top of the window or the bottom depending on the fragment
3100 coordinate origin convention (see TGSI_PROPERTY_FS_COORD_ORIGIN).
3101 The Z coordinate ranges from 0 to 1 to represent depth from the front
3102 to the back of the Z buffer. The W component contains the interpolated
3103 reciprocal of the vertex position W component (corresponding to gl_Fragcoord,
3104 but unlike d3d10 which interpolates the same 1/w but then gives back
3105 the reciprocal of the interpolated value).
3107 Fragment shaders may also declare an output register with
3108 TGSI_SEMANTIC_POSITION. Only the Z component is writable. This allows
3109 the fragment shader to change the fragment's Z position.
3116 For vertex shader outputs or fragment shader inputs/outputs, this
3117 label indicates that the register contains an R,G,B,A color.
3119 Several shader inputs/outputs may contain colors so the semantic index
3120 is used to distinguish them. For example, color[0] may be the diffuse
3121 color while color[1] may be the specular color.
3123 This label is needed so that the flat/smooth shading can be applied
3124 to the right interpolants during rasterization.
3128 TGSI_SEMANTIC_BCOLOR
3129 """"""""""""""""""""
3131 Back-facing colors are only used for back-facing polygons, and are only valid
3132 in vertex shader outputs. After rasterization, all polygons are front-facing
3133 and COLOR and BCOLOR end up occupying the same slots in the fragment shader,
3134 so all BCOLORs effectively become regular COLORs in the fragment shader.
3140 Vertex shader inputs and outputs and fragment shader inputs may be
3141 labeled with TGSI_SEMANTIC_FOG to indicate that the register contains
3142 a fog coordinate. Typically, the fragment shader will use the fog coordinate
3143 to compute a fog blend factor which is used to blend the normal fragment color
3144 with a constant fog color. But fog coord really is just an ordinary vec4
3145 register like regular semantics.
3151 Vertex shader input and output registers may be labeled with
3152 TGIS_SEMANTIC_PSIZE to indicate that the register contains a point size
3153 in the form (S, 0, 0, 1). The point size controls the width or diameter
3154 of points for rasterization. This label cannot be used in fragment
3157 When using this semantic, be sure to set the appropriate state in the
3158 :ref:`rasterizer` first.
3161 TGSI_SEMANTIC_TEXCOORD
3162 """"""""""""""""""""""
3164 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
3166 Vertex shader outputs and fragment shader inputs may be labeled with
3167 this semantic to make them replaceable by sprite coordinates via the
3168 sprite_coord_enable state in the :ref:`rasterizer`.
3169 The semantic index permitted with this semantic is limited to <= 7.
3171 If the driver does not support TEXCOORD, sprite coordinate replacement
3172 applies to inputs with the GENERIC semantic instead.
3174 The intended use case for this semantic is gl_TexCoord.
3177 TGSI_SEMANTIC_PCOORD
3178 """"""""""""""""""""
3180 Only available if PIPE_CAP_TGSI_TEXCOORD is exposed !
3182 Fragment shader inputs may be labeled with TGSI_SEMANTIC_PCOORD to indicate
3183 that the register contains sprite coordinates in the form (x, y, 0, 1), if
3184 the current primitive is a point and point sprites are enabled. Otherwise,
3185 the contents of the register are undefined.
3187 The intended use case for this semantic is gl_PointCoord.
3190 TGSI_SEMANTIC_GENERIC
3191 """""""""""""""""""""
3193 All vertex/fragment shader inputs/outputs not labeled with any other
3194 semantic label can be considered to be generic attributes. Typical
3195 uses of generic inputs/outputs are texcoords and user-defined values.
3198 TGSI_SEMANTIC_NORMAL
3199 """"""""""""""""""""
3201 Indicates that a vertex shader input is a normal vector. This is
3202 typically only used for legacy graphics APIs.
3208 This label applies to fragment shader inputs (or system values,
3209 depending on which one is supported by the driver) and indicates that
3210 the register contains front/back-face information.
3212 If it is an input, it will be a floating-point vector in the form (F, 0, 0, 1),
3213 where F will be positive when the fragment belongs to a front-facing polygon,
3214 and negative when the fragment belongs to a back-facing polygon.
3216 If it is a system value, it will be an integer vector in the form (F, 0, 0, 1),
3217 where F is 0xffffffff when the fragment belongs to a front-facing polygon and
3218 0 when the fragment belongs to a back-facing polygon.
3221 TGSI_SEMANTIC_EDGEFLAG
3222 """"""""""""""""""""""
3224 For vertex shaders, this sematic label indicates that an input or
3225 output is a boolean edge flag. The register layout is [F, x, x, x]
3226 where F is 0.0 or 1.0 and x = don't care. Normally, the vertex shader
3227 simply copies the edge flag input to the edgeflag output.
3229 Edge flags are used to control which lines or points are actually
3230 drawn when the polygon mode converts triangles/quads/polygons into
3234 TGSI_SEMANTIC_STENCIL
3235 """""""""""""""""""""
3237 For fragment shaders, this semantic label indicates that an output
3238 is a writable stencil reference value. Only the Y component is writable.
3239 This allows the fragment shader to change the fragments stencilref value.
3242 TGSI_SEMANTIC_VIEWPORT_INDEX
3243 """"""""""""""""""""""""""""
3245 For geometry shaders, this semantic label indicates that an output
3246 contains the index of the viewport (and scissor) to use.
3247 This is an integer value, and only the X component is used.
3249 If PIPE_CAP_TGSI_VS_LAYER_VIEWPORT or PIPE_CAP_TGSI_TES_LAYER_VIEWPORT is
3250 supported, then this semantic label can also be used in vertex or
3251 tessellation evaluation shaders, respectively. Only the value written in the
3252 last vertex processing stage is used.
3258 For geometry shaders, this semantic label indicates that an output
3259 contains the layer value to use for the color and depth/stencil surfaces.
3260 This is an integer value, and only the X component is used.
3261 (Also known as rendertarget array index.)
3263 If PIPE_CAP_TGSI_VS_LAYER_VIEWPORT or PIPE_CAP_TGSI_TES_LAYER_VIEWPORT is
3264 supported, then this semantic label can also be used in vertex or
3265 tessellation evaluation shaders, respectively. Only the value written in the
3266 last vertex processing stage is used.
3269 TGSI_SEMANTIC_CULLDIST
3270 """"""""""""""""""""""
3272 Used as distance to plane for performing application-defined culling
3273 of individual primitives against a plane. When components of vertex
3274 elements are given this label, these values are assumed to be a
3275 float32 signed distance to a plane. Primitives will be completely
3276 discarded if the plane distance for all of the vertices in the
3277 primitive are < 0. If a vertex has a cull distance of NaN, that
3278 vertex counts as "out" (as if its < 0);
3279 The limits on both clip and cull distances are bound
3280 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
3281 the maximum number of components that can be used to hold the
3282 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
3283 which specifies the maximum number of registers which can be
3284 annotated with those semantics.
3287 TGSI_SEMANTIC_CLIPDIST
3288 """"""""""""""""""""""
3290 Note this covers clipping and culling distances.
3292 When components of vertex elements are identified this way, these
3293 values are each assumed to be a float32 signed distance to a plane.
3296 Primitive setup only invokes rasterization on pixels for which
3297 the interpolated plane distances are >= 0.
3300 Primitives will be completely discarded if the plane distance
3301 for all of the vertices in the primitive are < 0.
3302 If a vertex has a cull distance of NaN, that vertex counts as "out"
3305 Multiple clip/cull planes can be implemented simultaneously, by
3306 annotating multiple components of one or more vertex elements with
3307 the above specified semantic.
3308 The limits on both clip and cull distances are bound
3309 by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_COUNT define which defines
3310 the maximum number of components that can be used to hold the
3311 distances and by the PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT
3312 which specifies the maximum number of registers which can be
3313 annotated with those semantics.
3314 The properties NUM_CLIPDIST_ENABLED and NUM_CULLDIST_ENABLED
3315 are used to divide up the 2 x vec4 space between clipping and culling.
3317 TGSI_SEMANTIC_SAMPLEID
3318 """"""""""""""""""""""
3320 For fragment shaders, this semantic label indicates that a system value
3321 contains the current sample id (i.e. gl_SampleID) as an unsigned int.
3322 Only the X component is used. If per-sample shading is not enabled,
3323 the result is (0, undef, undef, undef).
3325 TGSI_SEMANTIC_SAMPLEPOS
3326 """""""""""""""""""""""
3328 For fragment shaders, this semantic label indicates that a system
3329 value contains the current sample's position as float4(x, y, undef, undef)
3330 in the render target (i.e. gl_SamplePosition) when per-fragment shading
3331 is in effect. Position values are in the range [0, 1] where 0.5 is
3332 the center of the fragment.
3334 TGSI_SEMANTIC_SAMPLEMASK
3335 """"""""""""""""""""""""
3337 For fragment shaders, this semantic label can be applied to either a
3338 shader system value input or output.
3340 For a system value, the sample mask indicates the set of samples covered by
3341 the current primitive. If MSAA is not enabled, the value is (1, 0, 0, 0).
3343 For an output, the sample mask is used to disable further sample processing.
3345 For both, the register type is uint[4] but only the X component is used
3346 (i.e. gl_SampleMask[0]). Each bit corresponds to one sample position (up
3347 to 32x MSAA is supported).
3349 TGSI_SEMANTIC_INVOCATIONID
3350 """"""""""""""""""""""""""
3352 For geometry shaders, this semantic label indicates that a system value
3353 contains the current invocation id (i.e. gl_InvocationID).
3354 This is an integer value, and only the X component is used.
3356 TGSI_SEMANTIC_INSTANCEID
3357 """"""""""""""""""""""""
3359 For vertex shaders, this semantic label indicates that a system value contains
3360 the current instance id (i.e. gl_InstanceID). It does not include the base
3361 instance. This is an integer value, and only the X component is used.
3363 TGSI_SEMANTIC_VERTEXID
3364 """"""""""""""""""""""
3366 For vertex shaders, this semantic label indicates that a system value contains
3367 the current vertex id (i.e. gl_VertexID). It does (unlike in d3d10) include the
3368 base vertex. This is an integer value, and only the X component is used.
3370 TGSI_SEMANTIC_VERTEXID_NOBASE
3371 """""""""""""""""""""""""""""""
3373 For vertex shaders, this semantic label indicates that a system value contains
3374 the current vertex id without including the base vertex (this corresponds to
3375 d3d10 vertex id, so TGSI_SEMANTIC_VERTEXID_NOBASE + TGSI_SEMANTIC_BASEVERTEX
3376 == TGSI_SEMANTIC_VERTEXID). This is an integer value, and only the X component
3379 TGSI_SEMANTIC_BASEVERTEX
3380 """"""""""""""""""""""""
3382 For vertex shaders, this semantic label indicates that a system value contains
3383 the base vertex (i.e. gl_BaseVertex). Note that for non-indexed draw calls,
3384 this contains the first (or start) value instead.
3385 This is an integer value, and only the X component is used.
3387 TGSI_SEMANTIC_PRIMID
3388 """"""""""""""""""""
3390 For geometry and fragment shaders, this semantic label indicates the value
3391 contains the primitive id (i.e. gl_PrimitiveID). This is an integer value,
3392 and only the X component is used.
3393 FIXME: This right now can be either a ordinary input or a system value...
3399 For tessellation evaluation/control shaders, this semantic label indicates a
3400 generic per-patch attribute. Such semantics will not implicitly be per-vertex
3403 TGSI_SEMANTIC_TESSCOORD
3404 """""""""""""""""""""""
3406 For tessellation evaluation shaders, this semantic label indicates the
3407 coordinates of the vertex being processed. This is available in XYZ; W is
3410 TGSI_SEMANTIC_TESSOUTER
3411 """""""""""""""""""""""
3413 For tessellation evaluation/control shaders, this semantic label indicates the
3414 outer tessellation levels of the patch. Isoline tessellation will only have XY
3415 defined, triangle will have XYZ and quads will have XYZW defined. This
3416 corresponds to gl_TessLevelOuter.
3418 TGSI_SEMANTIC_TESSINNER
3419 """""""""""""""""""""""
3421 For tessellation evaluation/control shaders, this semantic label indicates the
3422 inner tessellation levels of the patch. The X value is only defined for
3423 triangle tessellation, while quads will have XY defined. This is entirely
3424 undefined for isoline tessellation.
3426 TGSI_SEMANTIC_VERTICESIN
3427 """"""""""""""""""""""""
3429 For tessellation evaluation/control shaders, this semantic label indicates the
3430 number of vertices provided in the input patch. Only the X value is defined.
3432 TGSI_SEMANTIC_HELPER_INVOCATION
3433 """""""""""""""""""""""""""""""
3435 For fragment shaders, this semantic indicates whether the current
3436 invocation is covered or not. Helper invocations are created in order
3437 to properly compute derivatives, however it may be desirable to skip
3438 some of the logic in those cases. See ``gl_HelperInvocation`` documentation.
3440 TGSI_SEMANTIC_BASEINSTANCE
3441 """"""""""""""""""""""""""
3443 For vertex shaders, the base instance argument supplied for this
3444 draw. This is an integer value, and only the X component is used.
3446 TGSI_SEMANTIC_DRAWID
3447 """"""""""""""""""""
3449 For vertex shaders, the zero-based index of the current draw in a
3450 ``glMultiDraw*`` invocation. This is an integer value, and only the X
3454 TGSI_SEMANTIC_WORK_DIM
3455 """"""""""""""""""""""
3457 For compute shaders started via opencl this retrieves the work_dim
3458 parameter to the clEnqueueNDRangeKernel call with which the shader
3462 TGSI_SEMANTIC_GRID_SIZE
3463 """""""""""""""""""""""
3465 For compute shaders, this semantic indicates the maximum (x, y, z) dimensions
3466 of a grid of thread blocks.
3469 TGSI_SEMANTIC_BLOCK_ID
3470 """"""""""""""""""""""
3472 For compute shaders, this semantic indicates the (x, y, z) coordinates of the
3473 current block inside of the grid.
3476 TGSI_SEMANTIC_BLOCK_SIZE
3477 """"""""""""""""""""""""
3479 For compute shaders, this semantic indicates the maximum (x, y, z) dimensions
3480 of a block in threads.
3483 TGSI_SEMANTIC_THREAD_ID
3484 """""""""""""""""""""""
3486 For compute shaders, this semantic indicates the (x, y, z) coordinates of the
3487 current thread inside of the block.
3490 TGSI_SEMANTIC_SUBGROUP_SIZE
3491 """""""""""""""""""""""""""
3493 This semantic indicates the subgroup size for the current invocation. This is
3494 an integer of at most 64, as it indicates the width of lanemasks. It does not
3495 depend on the number of invocations that are active.
3498 TGSI_SEMANTIC_SUBGROUP_INVOCATION
3499 """""""""""""""""""""""""""""""""
3501 The index of the current invocation within its subgroup.
3504 TGSI_SEMANTIC_SUBGROUP_EQ_MASK
3505 """"""""""""""""""""""""""""""
3507 A bit mask of ``bit index == TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3508 ``1 << subgroup_invocation`` in arbitrary precision arithmetic.
3511 TGSI_SEMANTIC_SUBGROUP_GE_MASK
3512 """"""""""""""""""""""""""""""
3514 A bit mask of ``bit index >= TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3515 ``((1 << (subgroup_size - subgroup_invocation)) - 1) << subgroup_invocation``
3516 in arbitrary precision arithmetic.
3519 TGSI_SEMANTIC_SUBGROUP_GT_MASK
3520 """"""""""""""""""""""""""""""
3522 A bit mask of ``bit index > TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3523 ``((1 << (subgroup_size - subgroup_invocation - 1)) - 1) << (subgroup_invocation + 1)``
3524 in arbitrary precision arithmetic.
3527 TGSI_SEMANTIC_SUBGROUP_LE_MASK
3528 """"""""""""""""""""""""""""""
3530 A bit mask of ``bit index <= TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3531 ``(1 << (subgroup_invocation + 1)) - 1`` in arbitrary precision arithmetic.
3534 TGSI_SEMANTIC_SUBGROUP_LT_MASK
3535 """"""""""""""""""""""""""""""
3537 A bit mask of ``bit index > TGSI_SEMANTIC_SUBGROUP_INVOCATION``, i.e.
3538 ``(1 << subgroup_invocation) - 1`` in arbitrary precision arithmetic.
3541 Declaration Interpolate
3542 ^^^^^^^^^^^^^^^^^^^^^^^
3544 This token is only valid for fragment shader INPUT declarations.
3546 The Interpolate field specifes the way input is being interpolated by
3547 the rasteriser and is one of TGSI_INTERPOLATE_*.
3549 The Location field specifies the location inside the pixel that the
3550 interpolation should be done at, one of ``TGSI_INTERPOLATE_LOC_*``. Note that
3551 when per-sample shading is enabled, the implementation may choose to
3552 interpolate at the sample irrespective of the Location field.
3554 The CylindricalWrap bitfield specifies which register components
3555 should be subject to cylindrical wrapping when interpolating by the
3556 rasteriser. If TGSI_CYLINDRICAL_WRAP_X is set to 1, the X component
3557 should be interpolated according to cylindrical wrapping rules.
3560 Declaration Sampler View
3561 ^^^^^^^^^^^^^^^^^^^^^^^^
3563 Follows Declaration token if file is TGSI_FILE_SAMPLER_VIEW.
3565 DCL SVIEW[#], resource, type(s)
3567 Declares a shader input sampler view and assigns it to a SVIEW[#]
3570 resource can be one of BUFFER, 1D, 2D, 3D, 1DArray and 2DArray.
3572 type must be 1 or 4 entries (if specifying on a per-component
3573 level) out of UNORM, SNORM, SINT, UINT and FLOAT.
3575 For TEX\* style texture sample opcodes (as opposed to SAMPLE\* opcodes
3576 which take an explicit SVIEW[#] source register), there may be optionally
3577 SVIEW[#] declarations. In this case, the SVIEW index is implied by the
3578 SAMP index, and there must be a corresponding SVIEW[#] declaration for
3579 each SAMP[#] declaration. Drivers are free to ignore this if they wish.
3580 But note in particular that some drivers need to know the sampler type
3581 (float/int/unsigned) in order to generate the correct code, so cases
3582 where integer textures are sampled, SVIEW[#] declarations should be
3585 NOTE: It is NOT legal to mix SAMPLE\* style opcodes and TEX\* opcodes
3588 Declaration Resource
3589 ^^^^^^^^^^^^^^^^^^^^
3591 Follows Declaration token if file is TGSI_FILE_RESOURCE.
3593 DCL RES[#], resource [, WR] [, RAW]
3595 Declares a shader input resource and assigns it to a RES[#]
3598 resource can be one of BUFFER, 1D, 2D, 3D, CUBE, 1DArray and
3601 If the RAW keyword is not specified, the texture data will be
3602 subject to conversion, swizzling and scaling as required to yield
3603 the specified data type from the physical data format of the bound
3606 If the RAW keyword is specified, no channel conversion will be
3607 performed: the values read for each of the channels (X,Y,Z,W) will
3608 correspond to consecutive words in the same order and format
3609 they're found in memory. No element-to-address conversion will be
3610 performed either: the value of the provided X coordinate will be
3611 interpreted in byte units instead of texel units. The result of
3612 accessing a misaligned address is undefined.
3614 Usage of the STORE opcode is only allowed if the WR (writable) flag
3619 ^^^^^^^^^^^^^^^^^^^^^^^^
3621 Properties are general directives that apply to the whole TGSI program.
3626 Specifies the fragment shader TGSI_SEMANTIC_POSITION coordinate origin.
3627 The default value is UPPER_LEFT.
3629 If UPPER_LEFT, the position will be (0,0) at the upper left corner and
3630 increase downward and rightward.
3631 If LOWER_LEFT, the position will be (0,0) at the lower left corner and
3632 increase upward and rightward.
3634 OpenGL defaults to LOWER_LEFT, and is configurable with the
3635 GL_ARB_fragment_coord_conventions extension.
3637 DirectX 9/10 use UPPER_LEFT.
3639 FS_COORD_PIXEL_CENTER
3640 """""""""""""""""""""
3642 Specifies the fragment shader TGSI_SEMANTIC_POSITION pixel center convention.
3643 The default value is HALF_INTEGER.
3645 If HALF_INTEGER, the fractionary part of the position will be 0.5
3646 If INTEGER, the fractionary part of the position will be 0.0
3648 Note that this does not affect the set of fragments generated by
3649 rasterization, which is instead controlled by half_pixel_center in the
3652 OpenGL defaults to HALF_INTEGER, and is configurable with the
3653 GL_ARB_fragment_coord_conventions extension.
3655 DirectX 9 uses INTEGER.
3656 DirectX 10 uses HALF_INTEGER.
3658 FS_COLOR0_WRITES_ALL_CBUFS
3659 """"""""""""""""""""""""""
3660 Specifies that writes to the fragment shader color 0 are replicated to all
3661 bound cbufs. This facilitates OpenGL's fragColor output vs fragData[0] where
3662 fragData is directed to a single color buffer, but fragColor is broadcast.
3665 """"""""""""""""""""""""""
3666 If this property is set on the program bound to the shader stage before the
3667 fragment shader, user clip planes should have no effect (be disabled) even if
3668 that shader does not write to any clip distance outputs and the rasterizer's
3669 clip_plane_enable is non-zero.
3670 This property is only supported by drivers that also support shader clip
3672 This is useful for APIs that don't have UCPs and where clip distances written
3673 by a shader cannot be disabled.
3678 Specifies the number of times a geometry shader should be executed for each
3679 input primitive. Each invocation will have a different
3680 TGSI_SEMANTIC_INVOCATIONID system value set. If not specified, assumed to
3683 VS_WINDOW_SPACE_POSITION
3684 """"""""""""""""""""""""""
3685 If this property is set on the vertex shader, the TGSI_SEMANTIC_POSITION output
3686 is assumed to contain window space coordinates.
3687 Division of X,Y,Z by W and the viewport transformation are disabled, and 1/W is
3688 directly taken from the 4-th component of the shader output.
3689 Naturally, clipping is not performed on window coordinates either.
3690 The effect of this property is undefined if a geometry or tessellation shader
3696 The number of vertices written by the tessellation control shader. This
3697 effectively defines the patch input size of the tessellation evaluation shader
3703 This sets the tessellation primitive mode, one of ``PIPE_PRIM_TRIANGLES``,
3704 ``PIPE_PRIM_QUADS``, or ``PIPE_PRIM_LINES``. (Unlike in GL, there is no
3705 separate isolines settings, the regular lines is assumed to mean isolines.)
3710 This sets the spacing mode of the tessellation generator, one of
3711 ``PIPE_TESS_SPACING_*``.
3716 This sets the vertex order to be clockwise if the value is 1, or
3717 counter-clockwise if set to 0.
3722 If set to a non-zero value, this turns on point mode for the tessellator,
3723 which means that points will be generated instead of primitives.
3725 NUM_CLIPDIST_ENABLED
3726 """"""""""""""""""""
3728 How many clip distance scalar outputs are enabled.
3730 NUM_CULLDIST_ENABLED
3731 """"""""""""""""""""
3733 How many cull distance scalar outputs are enabled.
3735 FS_EARLY_DEPTH_STENCIL
3736 """"""""""""""""""""""
3738 Whether depth test, stencil test, and occlusion query should run before
3739 the fragment shader (regardless of fragment shader side effects). Corresponds
3740 to GLSL early_fragment_tests.
3745 Which shader stage will MOST LIKELY follow after this shader when the shader
3746 is bound. This is only a hint to the driver and doesn't have to be precise.
3747 Only set for VS and TES.
3749 CS_FIXED_BLOCK_WIDTH / HEIGHT / DEPTH
3750 """""""""""""""""""""""""""""""""""""
3752 Threads per block in each dimension, if known at compile time. If the block size
3753 is known all three should be at least 1. If it is unknown they should all be set
3759 The MUL TGSI operation (FP32 multiplication) will return 0 if either
3760 of the operands are equal to 0. That means that 0 * Inf = 0. This
3761 should be set the same way for an entire pipeline. Note that this
3762 applies not only to the literal MUL TGSI opcode, but all FP32
3763 multiplications implied by other operations, such as MAD, FMA, DP2,
3764 DP3, DP4, DPH, DST, LOG, LRP, XPD, and possibly others. If there is a
3765 mismatch between shaders, then it is unspecified whether this behavior
3768 FS_POST_DEPTH_COVERAGE
3769 """"""""""""""""""""""
3771 When enabled, the input for TGSI_SEMANTIC_SAMPLEMASK will exclude samples
3772 that have failed the depth/stencil tests. This is only valid when
3773 FS_EARLY_DEPTH_STENCIL is also specified.
3776 Texture Sampling and Texture Formats
3777 ------------------------------------
3779 This table shows how texture image components are returned as (x,y,z,w) tuples
3780 by TGSI texture instructions, such as :opcode:`TEX`, :opcode:`TXD`, and
3781 :opcode:`TXP`. For reference, OpenGL and Direct3D conventions are shown as
3784 +--------------------+--------------+--------------------+--------------+
3785 | Texture Components | Gallium | OpenGL | Direct3D 9 |
3786 +====================+==============+====================+==============+
3787 | R | (r, 0, 0, 1) | (r, 0, 0, 1) | (r, 1, 1, 1) |
3788 +--------------------+--------------+--------------------+--------------+
3789 | RG | (r, g, 0, 1) | (r, g, 0, 1) | (r, g, 1, 1) |
3790 +--------------------+--------------+--------------------+--------------+
3791 | RGB | (r, g, b, 1) | (r, g, b, 1) | (r, g, b, 1) |
3792 +--------------------+--------------+--------------------+--------------+
3793 | RGBA | (r, g, b, a) | (r, g, b, a) | (r, g, b, a) |
3794 +--------------------+--------------+--------------------+--------------+
3795 | A | (0, 0, 0, a) | (0, 0, 0, a) | (0, 0, 0, a) |
3796 +--------------------+--------------+--------------------+--------------+
3797 | L | (l, l, l, 1) | (l, l, l, 1) | (l, l, l, 1) |
3798 +--------------------+--------------+--------------------+--------------+
3799 | LA | (l, l, l, a) | (l, l, l, a) | (l, l, l, a) |
3800 +--------------------+--------------+--------------------+--------------+
3801 | I | (i, i, i, i) | (i, i, i, i) | N/A |
3802 +--------------------+--------------+--------------------+--------------+
3803 | UV | XXX TBD | (0, 0, 0, 1) | (u, v, 1, 1) |
3804 | | | [#envmap-bumpmap]_ | |
3805 +--------------------+--------------+--------------------+--------------+
3806 | Z | XXX TBD | (z, z, z, 1) | (0, z, 0, 1) |
3807 | | | [#depth-tex-mode]_ | |
3808 +--------------------+--------------+--------------------+--------------+
3809 | S | (s, s, s, s) | unknown | unknown |
3810 +--------------------+--------------+--------------------+--------------+
3812 .. [#envmap-bumpmap] http://www.opengl.org/registry/specs/ATI/envmap_bumpmap.txt
3813 .. [#depth-tex-mode] the default is (z, z, z, 1) but may also be (0, 0, 0, z)
3814 or (z, z, z, z) depending on the value of GL_DEPTH_TEXTURE_MODE.