Reflection Info

This page describes the various values and extra information associated with the generated shader reflection.

Shader Base Types

Reflection uses a set of base types that can describe both numeric and opaque handle types. Any time a "base type" or "component type" is referenced in the reflection, it will take one of the following values:

Value	Type	Notes
1	Boolean	A 32-bit boolean type
2	Signed	Signed integer type
3	Unsigned	Unsigned integer type
4	Float	Floating point type
5	Sampler	A sampler type (combined image-sampler in Vulkan terms)
6	Image	A read/write image
7	ROBuffer	A read-only structured buffer
8	RWBuffer	A read-write structured buffer
9	ROTexels	A read-only texel buffer
10	RWTexels	A read-write texel buffer

Texel Rank

Texel-like objects will have a number of dimensions (known as the "rank") associated with them. They take the following values:

Value	Type	Notes
0	1D	1-dimensional
1	2D	2-dimensional
2	3D	3-dimensional
3	1D Array	Array of 1-dimensional objects
4	2D Array	Array of 2-dimensional objects
5	Cube	A cube-map
6	Buffer	Special rank for ROTexels and RWTexels

Numeric Types

Numeric types are associated with a base type (Signed, Unsigned, or Float), a size, and two dimension values.

The base types are described above.

The size is the component size of the numeric type. Right now, this will always be 4 as numerics of different sizes are not yet supported.

The first dimension value gives the vector size (or matrix row count), and the second dimension gives the matrix columns, or one. Examples:

Type	Base Type	Dim[0]	Dim[1]
float	Float	1	1
ivec2	Signed	2	1
float2x4	Float	4	2
float4x2	Float	2	4
float4x4	Float	4	4

Texel Format

Texel-like objects have texel format information associated with them. Texel formats are given as a tuple of format type, format size, and component count. The format type gives the base type of the texel, and can be one of:

Value	Name	Description
0	Signed	The texel components are a signed integer type
1	Unsigned	The texel components are an unsigned integer type
2	Float	The texel components are a floating point type
3	UNorm	The texel components are unsigned integers that are normalized to a floating point value in [0, 1]
4	SNorm	The texel components are signed integers that are normalized to a floating point value in [-1, 1]

All supported VSL formats and their associated format info is given here:

Format	Base Type	Type Size	Component Count
int4	Signed	4	4
int2	Signed	4	2
int	Signed	4	1
uint4	Unsigned	4	4
uint2	Unsigned	4	2
uint	Unsigned	4	1
float4	Float	4	4
float2	Float	4	2
float	Float	4	1
u8norm4	UNorm	1	4
u8norm2	UNorm	1	2
u8norm	UNorm	1	1
u16norm4	UNorm	2	4
u16norm2	UNorm	2	2
u16norm	UNorm	2	1
s8norm4	SNorm	1	4
s8norm2	SNorm	1	2
s8norm	SNorm	1	1
s16norm4	SNorm	2	4
s16norm2	SNorm	2	2
s16norm	SNorm	2	1