13-More Open GL-AppA 9/26/05 9:03 AM Page 1

Appendix A

Low-Level Shading
Languages Reference

Paul Baker

Instructions
Table A.1 lists all of the instructions that are common to both ARB_vertex_program and
ARB_fragment_program. Table A.2 lists the instructions that are only available in
ARB_vertex_program, and Table A.3 lists the instructions that are only available in
ARB_fragment_program.

Table A.1 Shared Instructions

Instruction Meaning Input Output Description
ABS Absolute Value (x, y, z, w) (|x|, |y|, |z|, |w|) Calculates the absolute value
of each component of the
source vector.
ADD Addition (x1, y1, z1, w1), (x1 + x2, y1 + y2, Adds the two source vectors
(x2, y2, z2, w2) z1 + z2, w1 + w2) component-wise.
DP3 3-component (x1, y1, z1, w1), (d, d, d, d) where Calculates a 3-component
dot product (x2, y2, z2, w2) + y1 * y2 + z1 * z2 scalar (dot) product of the
d = x1 * x2 source vectors and writes the
result to all components of
the destination vector.
DP4 4-component (x1, y1, z1, w1), (d, d, d, d) where Calculates a 4-component
dot product (x2, y2, z2, w2) d = x1 * x2 scalar (dot) product of the
+ y1 * y2 + z1 * z2 source vectors and writes
+ w1 * w2 the result to all components
of the destination vector.
continued
A1
13-More Open GL-AppA 9/26/05 9:03 AM Page 2

A2 Appendix A ■ Low-Level Shading Languages Reference

DPH Homogeneous (x1, y1, z1, w1), (d, d, d, d) where Calculates a 4-component
dot product (x2, y2, z2, w2) d = x1 * x2 + y1 scalar (dot) product of the
* y2 + z1 * z2 + w2 source vectors, with the
exception that w1 is forced to
1.0. Writes the result to all
components of the
destination vector.
DST Distance vector (x1, y1, z1, w1), (1.0, y1 * y2, z1, w2) When used with input vectors
(x2, y2, z2, w2) of the form:| (x1, y1, z1, w1)
= (N/A, d^2, d^2, N/A)
(x2, y2, z2, w2) = (N/A, 1/d,
N/A, 1/d) Calculates:
(1.0, d, d^2, 1/d)
EX2 Exponential s (2^s, 2^s, 2^s, 2^s) Calculates 2 raised to the
base 2 power of the scalar operand
and writes the result to all
four components of the
destination vector.
FLR Floor (x, y, z, w) (floor(x), floor(y), Calculates the floor of each
floor(z), floor(w)) component of the source
vector. The floor of a value is
the largest integer less than
or equal to the value.
FRC Fraction (x, y, z, w) (fraction(x), fraction (y), Calculates the fractional part
fraction (z), fraction (w)) of each component of the
source vector.
LG2 Logarithm s (log2(s), log2(s), Calculates the base 2
base 2 log2(s), log2(s)) logarithm of the scalar
operand and writes the result
to all four components of the
destination vector.
LIT Calculate (d, s, N/A, x) (1.0, D, S, 1.0) d is a diffuse dot product, s is
lighting a specular dot product, x is
coefficients the specular exponent, D is
the value to be multiplied by
the diffuse color, and S is the
value to be multiplied by the
specular color.
MAD Multiply and (x1, y1, z1, w1), (x1 * x2 + x3, y1 Multiplies the first two
accumulate (x2, y2, z2, w2), * y2 + y3, z1 * z2 + z3, source vectors and adds the
(x3, y3, z3, w3) w1 * w2 + w3) third source vector,
component-wise.
MAX Maximum (x1, y1, z1, w1), (max(x1, x2), max(y1, y2), Calculates the
(x2, y2, z2, w2) max(z1, z2), max(w1, w2)) component-wise maximum
of the two source vectors.
continued
13-More Open GL-AppA 9/26/05 9:03 AM Page 3

Instructions A3

MIN Minimum (x1, y1, z1, w1), (min(x1, x2), min(y1, y2), Calculates the component-
(x2, y2, z2, w2) min(z1, z2), min(w1, w2)) wise minimum of the two
source vectors.
MOV Move (x1, y1, z1, w1) (x1, y1, z1, w1) Copies the source vector into
the destination vector.
MUL Multiplication (x1, y1, z1, w1), (x1 * x2, y1 * y2, z1 Multiplies the two source
(x2, y2, z2, w2) * z2, w1 * w2) vectors component-wise.
POW Power s, p (s^p, s^p, s^p, s^p) Raises the first scalar
operand to the power of the
second scalar operand and
writes the result to all four
components of the
destination vector.
RCP Reciprocal s (1/s, 1/s, 1/s, 1/s) Calculates the reciprocal of
the scalar operand and
writes the result to all four
components of the
destination vector.
RSQ Reciprocal s (1/sqrt(s), 1/sqrt(s), Calculates the reciprocal
square root 1/sqrt(s), 1/sqrt(s)) of the scalar operand and
writes the result to all four
components of the
destination vector.
SGE Set On Greater (x1, y1, z1, w1), (x1 >= x2 ? 1.0 : 0.0, Compares the two source
Than Or Equal (x2, y2, z2, w2) y1 >= y2 ? 1.0 : 0.0, z1 vectors component-wise,
>= z2 ? 1.0 : 0.0, w1 setting the corresponding
component of the
destination vector if the first
source is greater than or
equal to the second.
SLT Set On (x1, y1, z1, w1), (x1 < x2 ? 1.0 : 0.0, y1 Compares the two source
Less Than (x2, y2, z2, w2) < y2 ? 1.0 : 0.0, z1 vectors component-wise,
< z2 ? 1.0 : 0.0, w1 setting the corresponding
< w2 ? 1.0 : 0.0) component of the destination
vector if the first source is
less than the second.
SUB Subtraction (x1, y1, z1, w1), (x1 –
x2, y1 –
y2, Subtracts the two source
(x2, y2, z2, w2) z1 –
z2, w1 –
w2) vectors component-wise.
continued
13-More Open GL-AppA 9/26/05 9:03 AM Page 4

A4 Appendix A ■ Low-Level Shading Languages Reference

SWZ Extended (x1, y1, z1, w1), (c1, c2, c3, c4) Reorders the components
Swizzle c1, c2, c3, c4 of a source vector, optionally
negating selected
components or replacing
them with a constant. The
parameters c1–c4 can be x,

x, y,
y, z,
z, w,
w, 0, 1
or
1. The corresponding
component of the source
vector, or the given constant,
is used as the component of
the destination vector.
XPD Cross Product (x1, y1, z1, N/A), (y1 * z2
z1 * y2, Calculates a vector (cross)
(x2, y2, z2, N/A) z1 * x2
x1 * z2, product using the first three
x1 * y2
y1 * x2, components of the source
UNDEF) vectors and writes the result
to the first three components
of the destination vector.

Table A.2 Vertex Program-Only Instructions

Instruction Meaning Input Output Description
ARL Address s floor(s) Loads an address register with a scalar.
Register Load
EXP Approximate s (2 ^ (floor(s)), Approximates 2 raised to the power of
Exponential s
floor(s), 2 the scalar operand and writes the
base 2 ^ s (approx), result to the z component of the result.
1.0) The x and y components of the result
can be used to generate a more
accurate approximation by calculating:
result.x * f(result.y) where f takes a
parameter x in [0.0, 1.0) and calculates
2^x.
LOG Approximate s (floor(log2(|s|)), Approximates the base-2 logarithm
Logarithm base 2 |s| / 2 ^ floor of the scalar operand and writes
(log2(|s|)), the result to the z component of the
log2(|s|) result. The x and y components of the
(approx), 1.0) result can be used to generate a more
accurate approximation by calculating:
result.x * f(result.y) where f takes a
parameter x in [1.0, 2.0) and
calculates 2^x.
13-More Open GL-AppA 9/26/05 9:03 AM Page 5

Instructions A5

Table A.3 Fragment Program-Only Instructions

Instruction Meaning Input Output Description
CMP Compare (x1, y1, z1, w1), (x1 < 0.0 ? x2 x3, Compares each component of
(x2, y2, z2, w2), y1 < 0.0 ? y2 : y3, the first source vector with zero,
(x3, y3, z3, w3) z1 < 0.0 ? z2 : z3, and selects the corresponding
w1 < 0.0 ? w2 : w3) component of the second or
third source vector depending on
the result.
COS Cosine s (cos(s), cos(s), Calculates the cosine of the
cos(s), cos(s)) scalar operand and writes the
result to all four components of
the destination vector.
KIL Kill fragment (x1, y1, z1, w1) None Prevents any further processing
of this fragment if any
component of the source vector
is less than zero.
LRP Linear (x1, y1, z1, w1), (x1 * x2 + (1.0
x1) Uses the first source vector as a
interpolation (x2, y2, z2, w2), * x3, y1 * y2 + blend factor to linearly
(x3, y3, z3, w3) (1.0
y1) * y3, z1 interpolate between the second
* z2 + (1.0
z1) and third source vectors,
* z3, w1 * w2 + component-wise.
(1.0 v w1) * w3)
SCS Sine/Cosine s (cos(s), sin(s), Calculates the cosine and sine of
UNDEF, UNDEF) the scalar operand and writes
the results to the x and y
components of the destination
vector respectively. The scalar
operand must lie in the range
[
, ].
SIN Sine s (sin(s), sin(s), Calculates the sine of the scalar
sin(s), sin(s)) operand and writes the result to
all four components of the
destination vector.
TEX Texture (s, t, r, q), TARGET Texture sample at (s, t, r).
Sample texture[n], Samples a texture using the
coordinates (s, t, r). TARGET may
be 1D, 2D, 3D, RECT or CUBE.
TXB Texture (s, t, r, q), TARGET Texture sample at (s, t, r),
Sample texture[n], with level-of-detail bias q. Samples
with Bias a texture using the coordinates
(s, t, r), using the q coordinate
as a level-of-detail bias.
TXP Texture (s, t, r, q), TARGET Texture sample at (s/q, t/q, r/q).
Sample texture[n], Samples a texture using the
with projective coordinates
Projection (s/q, t/q, r/q).
13-More Open GL-AppA 9/26/05 9:03 AM Page 6

A6 Appendix A ■ Low-Level Shading Languages Reference

Parameter Vectors
The OpenGL state variables listed in Table A.4 are available to vertex programs and frag-
ment programs as parameter vectors. Unless otherwise indicated, the variables are avail-
able to both types of program.

Table A.4 Parameter Vectors

Vector Description Values
state.material.ambient Front ambient material color (r, g, b, a)
state.material.diffuse Front diffuse material color (r, g, b, a)
state.material.specular Front specular material color (r, g, b, a)
state.material.emission Front emissive material color (r, g, b, a)
state.material.shininess Front material shininess (s, 0, 0, 1)

state.material.front.ambient Front ambient material color (r, g, b, a)

state.material.front.diffuse Front diffuse material color (r, g, b, a)
state.material.front.specular Front specular material color (r, g, b, a)
state.material.front.emission Front emissive material color (r, g, b, a)
state.material.front.shininess Front material shininess (s, 0, 0, 1)

state.material.back.ambient Back ambient material color (r, g, b, a)

state.material.back.diffuse Back diffuse material color (r, g, b, a)
state.material.back.specular Back specular material color (r, g, b, a)
state.material.back.emission Back emissive material color (r, g, b, a)
state.material.back.shininess Back material shininess (s, 0, 0, 1)

state.light[n].ambient Light n ambient color (r, g, b, a)

state.light[n].diffuse Light n diffuse color (r, g, b, a)
state.light[n].specular Light n specular color (r, g, b, a)
state.light[n].position Light n position (x, y, z, w)
state.light[n].attenuation Light n attenuation coefficients (a, b, c, e)
and spot light exponent
state.light[n].spot.direction Light n spot direction and cutoff (x, y, z, c)
angle cosine
state.light[n].half Light n infinite light/infinite (x, y, z, 1)
viewer half angle

state.lightmodel.ambient Light model ambient color (r, g, b, a)

state.lightmodel.scenecolor Light model front scene color (r, g, b, a)
state.lightmodel.front.scenecolor Light model front scene color (r, g, b, a)
continued
13-More Open GL-AppA 9/26/05 9:03 AM Page 7

Parameter Vectors A7

state.lightmodel.back.scenecolor Light model back scene color (r, g, b, a)

state.lightprod[n].ambient Light n ambient color * (r, g, b, a)

front material ambient color
state.lightprod[n].diffuse Light n diffuse color * (r, g, b, a)
front material diffuse color
state.lightprod[n].specular Light n specular color * (r, g, b, a)
front material specular color
state.lightprod[n].front.ambient Light n ambient color * (r, g, b, a)
front material ambient color
state.lightprod[n].front.diffuse Light n diffuse color * (r, g, b, a)
front material diffuse color
state.lightprod[n].front.specular Light n specular color * (r, g, b, a)
front material specular color

state.lightprod[n].back.ambient Light n ambient color * (r, g, b, a)

back material ambient color
state.lightprod[n].back diffuse Light n diffuse color * (r, g, b, a)
back material diffuse color
state.lightprod[n].back specular Light n specular color * (r, g, b, a)
back material specular color

state.texgen[n].eye.s TexGen eye linear plane (a, b, c, d) *

coefficients, unit n, s coordinate
state.texgen[n].eye.t TexGen eye linear plane coefficients, (a, b, c, d) *
unit n, t coordinate
state.texgen[n].eye.r TexGen eye linear plane coefficients, (a, b, c, d) *
unit n, r coordinate
state.texgen[n].eye.q TexGen eye linear plane coefficients, (a, b, c, d) *
unit n, q coordinate
state.texgen[n].object.s TexGen object linear plane coefficients, (a, b, c, d) *
unit n, s coordinate
state.texgen[n].object.t TexGen object linear plane coefficients, (a, b, c, d) *
unit n, t coordinate
state.texgen[n].object.r TexGen object linear plane coefficients, (a, b, c, d) *
unit n, r coordinate
state.texgen[n].object.q TexGen object linear plane coefficients, (a, b, c, d) *
unit n, q coordinate

state.texenv[n].color Texture environment color, unit n (r, g, b, a) †

state.fog.color Fog color (r, g, b, a)

continued
13-More Open GL-AppA 9/26/05 9:03 AM Page 8

A8 Appendix A ■ Low-Level Shading Languages Reference

state.fog.params Fog density, linear start and end, (d, s, e, r)

and 1/(end-start)

state.depth.range Depth range near, far and (far-near) (n, f, d, 1) †

state.clip[n].plane Clip plane n coefficients (a, b, c, d) *

state.point.size Point size, minimum and maximum (s, n, x, f) *

size clamps and fade threshold
state.point.attenuation Point-size attenuation coefficients (a, b, c, 1) *

state.matrix.modelview Modelview matrix ‡

state.matrix.modelview[n] Modelview matrix n ‡
state.matrix.projection Projection matrix ‡
state.matrix.mvp Modelview-projection matrix ‡
state.matrix.texture[n] Texture matrix n ‡
state.matrix.palette[n] Modelview palette matrix n ‡
state.matrix.program[n] Generic program matrix n ‡
* Only available in ARB_vertex_program
† Only available in ARB_fragment_program
‡ The matrix parameters fill four of the standard 4-component vectors by default, each holding one
row of the 44 matrix. A contiguous subset of the rows can be selected by appending ".row[x..y]" to
the state variable name. You can also append ".inverse", ".transpose", or ".invtrans" to the matrix
parameters in order to receive the inverse, transpose, or inverse transpose, respectively, of the
requested matrix. For example, in order to have rows 1 and 2 of the inverse transpose modelview
matrix placed in mvit[0] and mvit[1] respectively, the following command is used:
PARAM mvit[] = {state.matrix.modelview.invtrans.row[1..2] };
13-More Open GL-AppA 9/26/05 9:03 AM Page 9

ARB_vertex_program Result Vectors A9

ARB_vertex_program Attribute Vectors

Table A.5 lists the vertex attributes that are available.

Table A.5 Vertex Program Attribute Vectors

Vector Description Values
vertex.position Object space position (x, y, z, w)
vertex.normal Normal (x, y, z, 1)
vertex.color Primary color (r, g, b, a)
vertex.color.primary Primary color (r, g, b, a)
vertex.color.secondary Secondary color (r, g, b, a)
vertex.texcoord Texture coordinate set 0 (s, t, r, q)
vertex.texcoord[n] Texture coordinate set n (s, t, r, q)
vertex.fogcoord Fog coordinate (f, 0, 0, 1)
vertex.weight Vertex weights 0–3 (w, w, w, w)
vertex.weight[n] Vertex weights n–n+3 (w, w, w, w)
vertex.matrixindex Vertex matrix indices 0–3 (w, w, w, w)
vertex.matrixindex[n] Vertex matrix indices n–n+3 (w, w, w, w)
vertex.attrib[n] Vertex generic attribute n (x, y, z, w)

ARB_vertex_program Result Vectors

The results that may be written by a vertex program are shown in Table A.6.

Table A.6 Vertex Program Result Vectors

Vector Description Values
result.position Clip-space position (x, y, z, w)
result.color Front primary color (r, g, b, a)
result.color.primary Front primary color (r, g, b, a)
result.color.secondary Front secondary color (r, g, b, a)
result.color.front Front primary color (r, g, b, a)
result.color.front.primary Front primary color (r, g, b, a)
result.color.front.secondary Front secondary color (r, g, b, a)
result.color.back Back primary color (r, g, b, a)
result.color.back.primary Back primary color (r, g, b, a)
result.color.back.secondary Back secondary color (r, g, b, a)
result.texcoord Texture coordinate set 0 (s, t, r, q)
result.texcoord[n] Texture coordinate set n (s, t, r, q)
result.fogcoord Fog coordinate (f, N/A, N/A, N/A)
result.pointsize Point size (s, N/A, N/A, N/A)
13-More Open GL-AppA 9/26/05 9:03 AM Page 10

A10 Appendix A ■ Low-Level Shading Languages Reference

ARB_fragment_program Attribute Vectors

The available attributes of a fragment are listed in Table A.7.

Table A.7 Fragment Program Attribute Vectors

Vector Description Values
fragment.position Window position (x, y, z, 1/w)
fragment.color Primary color (r, g, b, a)
fragment.color.primary Primary color (r, g, b, a)
fragment.color.secondary Secondary color (r, g, b, a)
fragment.texcoord Texture coordinate set 0 (s, t, r, q)
fragment.texcoord[n] Texture coordinate set n (s, t, r, q)
fragment.fogcoord Fog coordinate (f, 0, 0, 1)

ARB_fragment_program Result Vectors

Table A.8 lists the results which may be written by a fragment program.

Table A.8 Fragment Program Result Vectors

Vector Description Values
result.color Final color (r, g, b, a)
result.color[n] Final color (when using multiple render buffers) (r, g, b, a)
result.depth Final depth (N/A, N/A, d, N/A)