Vulkan-Docs/doc/specs/vulkan/chapters/interfaces.txt

1181 lines
53 KiB
Plaintext
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

// Copyright (c) 2015-2016 The Khronos Group Inc.
// Copyright notice at https://www.khronos.org/registry/speccopyright.html
[[interfaces]]
= Shader Interfaces
When a pipeline is created, the set of shaders specified in the
corresponding stext:Vk*PipelineCreateInfo structure are implicitly linked
at a number of different interfaces.
* <<interfaces-iointerfaces,Shader Input and Output Interface>>
* <<interfaces-vertexinput,Vertex Input Interface>>
* <<interfaces-fragmentoutput,Fragment Output Interface>>
* <<interfaces-inputattachment,Fragment Input Attachment Interface>>
* <<interfaces-resources,Shader Resource Interface>>
[[interfaces-iointerfaces]]
== Shader Input and Output Interfaces
When multiple stages are present in a pipeline, the outputs of one
stage form an interface with the inputs of the next stage. When such an
interface involves a shader, shader outputs are matched against the inputs
of the next stage, and shader inputs are matched against the outputs of the
previous stage.
There are two classes of variables that can: be matched between shader
stages, built-in variables and user-defined variables. Each class has a
different set of matching criteria. Generally, when non-shader stages are
between shader stages, the user-defined variables, and most built-in
variables, form an interface between the shader stages.
The variables forming the input or output _interfaces_ are listed as
operands to the code:OpEntryPoint instruction and are declared with the
code:Input or code:Output storage classes, respectively, in the SPIR-V
module.
[[interfaces-iointerfaces-builtin]]
=== Built-in Interface Block
Shader <<interfaces-builtin-variables,built-in>> variables meeting the
following requirements define the _built-in interface block_. They must: be:
* explicitly declared (there are no implicit built-ins),
* identified with a code:BuiltIn decoration,
* form object types as described in the
<<interfaces-builtin-variables,Built-in Variables>> section, and
* declared in a block whose top-level members are the built-ins.
Built-ins only participate in interface matching if they are declared
in such a block. They mustnot: have any code:Location or code:Component
decorations.
There must: be no more than one built-in interface block per shader per
interface.
[[interfaces-iointerfaces-user]]
=== User-defined Variable Interface
The remaining variables listed by code:OpEntryPoint with the code:Input or
code:Output storage class form the _user-defined variable interface_.
These variables must: be identified with a code:Location decoration and
can: also be identified with a code:Component decoration.
[[interfaces-iointerfaces-matching]]
=== Interface Matching
A user-defined output variable is considered
to match an input variable in the subsequent stage if the two variables
are declared with the same code:Location and code:Component decoration
and match in type and decoration, except that
<<shaders-interpolation-decorations,interpolation decorations>> are not
required: to match. For the purposes of interface
matching, variables declared without a code:Component decoration are
considered to have a code:Component decoration of zero.
Variables or block members declared as structures are considered to match
in type if and only if the structure members match in type, decoration,
number, and declaration order. Variables or block members declared as arrays
are considered to match in type only if both declarations specify the
same element type and size.
Tessellation control shader per-vertex output variables and blocks, and
tessellation control, tessellation evaluation, and geometry shader
per-vertex input variables and blocks are required to be declared as
arrays, with each element representing input or output values for a
single vertex of a multi-vertex primitive. For the purposes of interface
matching, the outermost array dimension of such variables and blocks
is ignored.
At an interface between two non-fragment shader stages, the built-in
interface block must: match exactly, as described above. At an interface
involving the fragment shader inputs, the presence or absence of any
built-in output does not affect the interface matching.
Any input value to a shader stage is well-defined as long as the preceding
stages writes to a matching output, as described above.
Additionally, scalar and vector inputs are well-defined if there
is a corresponding output satisfying all of the following conditions:
* the input and output match exactly in decoration,
* the output is a vector with the same basic type and has
at least as many components as the input, and
* the common component type of the input and output is
32-bit integer or floating-point (64-bit component types are excluded).
In this case, the components of the input will be taken from the first
components of the output, and any extra components of the output
will be ignored.
[[interfaces-iointerfaces-locations]]
=== Location Assignment
This section describes how many locations are consumed by a given
type. As mentioned above, geometry shader inputs, tessellation control
shader inputs and outputs, and tessellation evaluation inputs all have
an additional level of arrayness relative to other shader inputs and
outputs. This outer array level is removed from the type before considering
how many locations the type consumes.
The code:Location value specifies an interface slot comprised of a
32-bit four-component vector conveyed between stages. The code:Component
specifies <<interfaces-iointerfaces-components,components>> within
these vector locations. Only types with widths of 32 or 64 are supported in
shader interfaces.
Inputs and outputs of the following types consume a single interface
location:
* 32-bit scalar and vector types, and
* 64-bit scalar and 2-component vector types.
64-bit three- and four-component vectors consume two consecutive locations.
If a declared input or output is an array of size _n_ and each element
takes _m_ locations, it will be assigned _m_ × _n_ consecutive locations
starting with the location specified.
If the declared input or output is an _n_ × _m_ 32- or 64-bit matrix,
it will be assigned multiple locations starting with the location specified.
The number of locations assigned for each matrix will be the same
as for an _n_-element array of _m_-component vectors.
The layout of a structure type used as an code:Input or code:Output depends
on whether it is also a code:Block (i.e. has a code:Block decoration).
If it is a not a code:Block, then the structure type must: have a
code:Location decoration. Its members are assigned consecutive locations
in their declaration order, with the first member assigned to the
location specified for the structure type. The members, and their nested
types, mustnot: themselves have code:Location decorations.
If the structure type is a code:Block but without a code:Location, then
each of its members must: have a code:Location decoration. If it is a
code:Block with a code:Location decoration, then its first member is
assigned to the location specified for the code:Block, any member with
its own code:Location decoration is assigned that location, and otherwise
each subsequent member is assigned consecutive locations in declaration
order.
The locations consumed by block and structure members are determined
by applying the rules above in a depth-first traversal of the instantiated
members as though the structure or block member were declared as an input or
output variable of the same type.
Any two inputs listed as operands on the same code:OpEntryPoint mustnot: be
assigned the same location, either explicitly or implicitly.
Any two outputs listed as operands on the same code:OpEntryPoint mustnot:
be assigned the same location, either explicitly or implicitly.
The number of input and output locations available for a shader input
or output interface are limited, and dependent on the shader stage
as described in <<interfaces-iointerfaces-limits>>.
[[interfaces-iointerfaces-limits]]
.Shader Input and Output Locations
[width="90%",cols="<6,<13",options="header"]
|=============================
| Shader Interface | Locations Available
| vertex input | pname:maxVertexInputAttributes
| vertex output | pname:maxVertexOutputComponents / 4
| tessellation control input | pname:maxTessellationControlPerVertexInputComponents / 4
| tessellation control output | pname:maxTessellationControlPerVertexOutputComponents / 4
| tessellation evaluation input | pname:maxTessellationEvaluationInputComponents / 4
| tessellation evaluation output| pname:maxTessellationEvaluationOutputComponents / 4
| geometry input | pname:maxGeometryInputComponents / 4
| geometry output | pname:maxGeometryOutputComponents / 4
| fragment input | pname:maxFragmentInputComponents / 4
| fragment output | pname:maxFragmentOutputAttachments
|=============================
[[interfaces-iointerfaces-components]]
=== Component Assignment
The code:Component decoration allows the code:Location to be more
finely specified for scalars and vectors, down to the individual
components within a location that are consumed.
The components within a location are 0, 1, 2, and 3.
A variable or block member starting at component N
will consume components N, N+1, N+2, ... up through its size.
For single precision types, it is invalid if this sequence of
components gets larger than 3. A scalar 64-bit type will consume
two of these components in sequence, and a
two-component 64-bit vector type will consume all four components
available within a location. A three- or four-component 64-bit vector
type mustnot: specify a code:Component decoration. A three-component
64-bit vector type will consume all four components of the first location
and components 0 and 1 of the second location. This leaves components
2 and 3 available for other component-qualified declarations.
A scalar or two-component 64-bit data type mustnot: specify a
code:Component decoration of 1 or 3.
A code:Component decoration mustnot: be specified for any type that is
not a scalar or vector.
[[interfaces-vertexinput]]
== Vertex Input Interface
When the vertex stage is present in a pipeline, the vertex shader input
variables form an interface with the vertex input attributes. The vertex
shader input variables are matched by the code:Location and
code:Component decorations to the vertex input attributes specified
in the pname:pVertexInputState member of the
slink:VkGraphicsPipelineCreateInfo structure.
The vertex shader input variables listed by code:OpEntryPoint with the
code:Input storage class form the _vertex input interface_. These variables
must: be identified with a code:Location decoration and can: also be
identified with a code:Component decoration.
For the purposes of interface
matching: variables declared without a code:Component decoration
are considered to have a code:Component decoration of zero.
The number of available vertex input locations is given by the
pname:maxVertexInputAttributes member of the sname:VkPhysicalDeviceLimits
structure.
See <<fxvertex-attrib-location>> for details.
All vertex shader inputs declared as above must: have a corresponding
attribute and binding in the pipeline.
[[interfaces-fragmentoutput]]
== Fragment Output Interface
When the fragment stage is present in a pipeline, the fragment shader
outputs form an interface with the output attachments of the current
subpass. The fragment shader output variables are matched by the
code:Location and code:Component decorations to the color attachments
specified in the pname:pColorAttachments array of the
slink:VkSubpassDescription structure that describes the subpass that the
fragment shader is executed in.
The fragment shader output variables listed by code:OpEntryPoint with the
code:Output storage class form the _fragment output interface_.
These variables must: be identified with a code:Location decoration.
They can: also be identified with a code:Component decoration and/or
an code:Index decoration. For the
purposes of interface matching: variables declared without a code:Component
decoration are considered to have a code:Component decoration of zero,
and variables declared without an code:Index decoration are considered
to have an code:Index decoration of zero.
A fragment shader output variable identified with a code:Location decoration
of _i_ is directed to the color attachment indicated by
pname:pColorAttachments[_i_], after passing through the blending unit as
described in <<framebuffer-blending>>, if enabled. Locations are consumed as
described in <<interfaces-iointerfaces-locations,Location Assignment>>. The
number of available fragment output locations is given by the
pname:maxFragmentOutputAttachments member of the
sname:VkPhysicalDeviceLimits structure.
Components of the output variables are assigned as described in
<<interfaces-iointerfaces-components,Component Assignment>>.
Output components identified as 0, 1, 2, and 3 will be directed
to the R, G, B, and A inputs to the blending unit, respectively,
or to the output attachment if blending is disabled.
If two variables are placed within the same location, they must
have the same underlying type (floating-point or integer).
Fragment outputs identified with an code:Index of zero are directed
to the first input of the blending unit associated with the
corresponding code:Location. Outputs identified with an code:Index
of one are directed to the second input of the corresponding
blending unit.
No _component aliasing_ of output variables is allowed, that is
there mustnot: be two output variables which have the same location,
component, and index, either explicitly declared or implied.
Output values written by a fragment shader must: be declared with
either code:OpTypeFloat or code:OpTypeInt, and a Width of 32.
Composites of these types are also permitted. If the color attachment has a
signed or unsigned normalized fixed-point format, color values are assumed
to be floating-point and are converted to fixed-point as described in
<<fundamentals-fixedfpconv>>; otherwise no type conversion
is applied. If the type of the values written by the fragment shader do
not match the format of the corresponding color attachment, the result is
undefined for those components.
[[interfaces-inputattachment]]
== Fragment Input Attachment Interface
When a fragment stage is present in a pipeline, the fragment shader
subpass inputs form an interface with the input attachments of the
current subpass. The fragment shader subpass input variables are
matched by code:InputAttachmentIndex decorations to the input
attachments specified in the pname:pInputAttachments array of the
slink:VkSubpassDescription structure that describes the subpass that
the fragment shader is executed in.
The fragment shader input variables listed by code:OpEntryPoint with the
code:Input storage class and a decoration of code:InputAttachmentIndex
form the _fragment input attachment interface_. These variables must:
be declared with a type of code:OpTypeImage and a code:Dim operand
of code:SubpassData.
A fragment shader input variable identified with an code:InputAttachmentIndex
decoration of _i_ reads from the input attachment indicated by
pname:pInputAttachments[_i_]. If the input variable is declared as
an array of size N, it consumes N consecutive input attachments, starting
with the index specified. There mustnot: be more than one input variable
with the same code:InputAttachmentIndex whether explicitly declared or
implied by an array declaration. The number of available input attachment
indices is given by the pname:maxPerStageDescriptorInputAttachments member
of the sname:VkPhysicalDeviceLimits structure.
Variables identified with the code:InputAttachmentIndex must: only be
used by a fragment stage. The basic data type (floating-point,
integer, unsigned integer) of the subpass input must: match the basic
format of the corresponding input attachment, or the values of subpass
loads from these variables are undefined.
See <<descriptorsets-inputattachment>> for more details.
[[interfaces-resources]]
== Shader Resource Interface
When a shader stage accesses buffer or image resources, as described
in the <<descriptorsets,Resource Descriptors>> section, the shader
resource variables must be matched with the
<<descriptorsets-pipelinelayout,pipeline layout>> that is provided
at pipeline creation time.
The set of shader resources that form the _shader resource interface_
for a stage are the variables statically used by code:OpEntryPoint
with the storage classes of code:Uniform, code:UniformConstant, and
code:PushConstant. For the fragment shader, the variables identified by
operands to code:OpEntryPoint with a storage class of code:Input and a
decoration of code:InputAttachmentIndex are also included in this interface.
The shader resource interface can be further broken down into two
sub-interfaces: the push constant interface and the descriptor
set interface.
[[interfaces-resources-pushconst]]
=== Push Constant Interface
The shader variables defined with a storage class of code:PushConstant that
are statically used by the shader entry points for the pipeline
define the _push constant interface_. They must: be:
* typed as code:OpTypeStruct,
* identified with a code:Block decoration, and
* laid out explicitly using the code:Offset, code:ArrayStride, and
code:MatrixStride decorations as specified in
<<interfaces-resources-layout,Offset and Stride Assignment>>.
There must: be no more than one push constant block statically used per
shader entry point.
Each variable in a push constant block must: be placed at an code:Offset
such that the entire constant value is entirely contained within the
slink:VkPushConstantRange for each code:OpEntryPoint that uses it, and the
pname:stageFlags for that range must: specify the appropriate
elink:VkShaderStageFlagBits for that stage. The code:Offset decoration for
any variable in a push constant block mustnot: cause the space required for
that variable to extend outside the range latexmath:[$[0,
\mathit{maxPushConstantsSize})$].
Any variable in a push constant block that is declared as an array must:
only be accessed with dynamically uniform indices.
[[interfaces-resources-descset]]
=== Descriptor Set Interface
The _descriptor set interface_ is comprised of the shader variables with the
storage classes of code:Uniform, code:UniformConstant, and the variables in
the <<interfaces-inputattachment,fragment input attachment interface>>,
that are statically used by the shader entry points for the pipeline.
These variables must: have code:DescriptorSet and code:Binding decorations
specified, which are assigned and matched with the
sname:VkDescriptorSetLayout objects in the pipeline layout as described in
<<interfaces-resources-setandbinding,DescriptorSet and Binding Assignment>>.
Variables identified with the code:UniformConstant storage class are used
only as handles to refer to opaque resources. Such variables must: be typed
as code:OpTypeImage, code:OpTypeSampler, code:OpTypeSampledImage, or arrays
of only these types. Variables of type code:OpTypeImage must: have a
code:Sampled operand of 1 (sampled image) or 2 (storage image).
Any array of these types must: only be indexed with constant integral
expressions, except under the following conditions:
* For arrays of code:OpTypeImage variables with code:Sampled operand of 2,
if the pname:shaderStorageImageArrayDynamicIndexing feature is enabled
and the shader module declares the code:StorageImageArrayDynamicIndexing
capability, the array must: only be indexed by dynamically uniform
expressions.
* For arrays of code:OpTypeSampler, code:OpTypeSampledImage variables, or
code:OpTypeImage variables with code:Sampled operand of 1,
if the pname:shaderSampledImageArrayDynamicIndexing feature is enabled
and the shader module declares the code:SampledImageArrayDynamicIndexing
capability, the array must: only be indexed by dynamically uniform
expressions.
The code:Sampled code:Type of an code:OpTypeImage declaration must: match
the same basic data type as the corresponding resource, or the values
obtained by reading or sampling from this image are undefined.
The code:Image code:Format of an code:OpTypeImage declaration mustnot: be
*Unknown*, for variables which are used for code:OpImageRead or
code:OpImageWrite operations, except under the following conditions:
* For code:OpImageWrite, if the pname:shaderStorageImageWriteWithoutFormat
feature is enabled and the shader module declares the
code:StorageImageWriteWithoutFormat capability.
* For code:OpImageRead, if the pname:shaderStorageImageReadWithoutFormat
feature is enabled and the shader module declares the
code:StorageImageReadWithoutFormat capability.
Variables identified with the code:Uniform storage class are used to access
transparent buffer backed resources. Such variables must: be:
* typed as code:OpTypeStruct, or arrays of only this type,
* identified with a code:Block or code:BufferBlock decoration, and
* laid out explicitly using the code:Offset, code:ArrayStride, and
code:MatrixStride decorations as specified in
<<interfaces-resources-layout,Offset and Stride Assignment>>.
Any array of these types must: only be indexed with constant integral
expressions, except under the following conditions.
* For arrays of code:Block variables, if the
pname:shaderUniformBufferArrayDynamicIndexing feature is enabled and
the shader module declares the code:UniformBufferArrayDynamicIndexing
capability, the array must: only be indexed by dynamically uniform
expressions.
* For arrays of code:BufferBlock variables, if the
pname:shaderStorageBufferArrayDynamicIndexing feature is enabled and
the shader module declares the code:StorageBufferArrayDynamicIndexing
capability, the array must: only be indexed by dynamically uniform
expressions.
The code:Offset decoration for any variable in a code:Block mustnot:
cause the space required for that variable to extend outside the
range latexmath:[$[0, \mathit{maxUniformBufferRange})$]. The code:Offset
decoration for any variable in a code:BufferBlock mustnot: cause the
space required for that variable to extend outside the range
latexmath:[$[0, \mathit{maxStorageBufferRange})$].
Variables identified with a storage class of code:Input and a decoration of
code:InputAttachmentIndex must be declared as described
<<interfaces-inputattachment,above>>.
Each shader variable declaration must: refer to the same type of resource as
is indicated by the pname:descriptorType. See
<<interfaces-resources-correspondence,Shader Resource and Descriptor Type
Correspondence>> for the relationship between shader declarations and
descriptor types.
[[interfaces-resources-correspondence]]
.Shader Resource and Descriptor Type Correspondence
[width="90%",cols="<1,<2",options="header"]
|=============================
| Resource type | Descriptor Type
| sampler |VK_DESCRIPTOR_TYPE_SAMPLER
| sampled image |VK_DESCRIPTOR_TYPE_SAMPLED_IMAGE
| storage image |VK_DESCRIPTOR_TYPE_STORAGE_IMAGE
| combined image sampler | VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER
| uniform texel buffer | VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER
| storage texel buffer | VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER
| uniform buffer | VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER +
VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER_DYNAMIC
| storage buffer | VK_DESCRIPTOR_TYPE_STORAGE_BUFFER +
VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC
| input attachment| VK_DESCRIPTOR_TYPE_INPUT_ATTACHMENT
|=============================
.Shader Resource and Storage Class Correspondence
[width="100%",cols="<21%,<22%,<27%,<30%",options="header"]
|=============================
| Resource type | Storage Class | Type | Decoration(s)^1^
| sampler
| code:UniformConstant | code:OpTypeSampler |
| sampled image
| code:UniformConstant | code:OpTypeImage (code:Sampled=1)|
| storage image
| code:UniformConstant | code:OpTypeImage (code:Sampled=2) |
| combined image sampler
| code:UniformConstant | code:OpTypeSampledImage |
| uniform texel buffer
| code:UniformConstant | code:OpTypeImage (code:Dim=code:Buffer, code:Sampled=1) |
| storage texel buffer
| code:UniformConstant | code:OpTypeImage (code:Dim=code:Buffer, code:Sampled=2) |
| uniform buffer
| code:Uniform | code:OpTypeStruct
| code:Block, code:Offset, (code:ArrayStride), (code:MatrixStride)
| storage buffer
| code:Uniform | code:OpTypeStruct
| code:BufferBlock, code:Offset, (code:ArrayStride), (code:MatrixStride)
| input attachment
| code:Input | code:OpTypeImage (code:Dim=code:SubpassData) | code:InputAttachmentIndex
|=============================
1:: in addition to code:DescriptorSet and code:Binding
[[interfaces-resources-setandbinding]]
=== DescriptorSet and Binding Assignment
A variable identified with a code:DescriptorSet decoration of
latexmath:[$s$] and a code:Binding decoration of latexmath:[$b$] indicates
that this variable is associated with the slink:VkDescriptorSetLayoutBinding
that has a pname:binding equal to latexmath:[$b$] in pname:pSetLayouts[_s_]
that was specified in slink:VkPipelineLayoutCreateInfo.
The range of descriptor sets is between zero and
pname:maxBoundDescriptorSets minus one, inclusive. If a descriptor set value
is statically used by an entry point there must: be an associated
pname:pSetLayout in the corresponding pipeline layout as described in
<<descriptorsets-pipelinelayout-consistency,Pipeline Layouts consistency>>.
If the code:Binding decoration is used with an array, the entire array is
identified with that binding value. The size of the array declaration must:
be no larger than the pname:descriptorCount of that
sname:VkDescriptorSetLayoutBinding. The index of each element of the array
is referred to as the _arrayElement_. For the purposes of interface matching
and descriptor set <<descriptorsets-updates,operations>>, if a resource
variable is not an array, it is treated as if it has an arrayElement of
zero.
The binding can be any 32-bit unsigned integer value, as described in
<<descriptorsets-setlayout>>. Each descriptor set has its own binding
name space.
There is a limit on the number of resources of each type that can: be
accessed by a pipeline stage as shown in
<<interfaces-resources-limits,Shader Resource Limits>>.
The ``Resources Per Stage'' column gives the limit on the number each type
of resource that can: be statically used for an entry point in any given
stage in a pipeline. The ``Resource Types'' column lists which resource
types are counted against the limit. Some resource types count against
multiple limits.
If multiple entry points in the same pipeline refer to the same set and
binding, all variable definitions with that code:DescriptorSet and
code:Binding must have the same basic type.
Not all descriptor sets and bindings specified in a pipeline layout need to
be used in a particular shader stage or pipeline, but if a
code:DescriptorSet and code:Binding decoration is specified for a variable
that is statically used in that shader there must: be a pipeline layout
entry identified with that descriptor set and pname:binding and the
corresponding pname:stageFlags must: specify the appropriate
elink:VkShaderStageFlagBits for that stage.
[[interfaces-resources-limits]]
.Shader Resource Limits
[width="80%",cols="<35,<23",options="header"]
|=============================
| Resources per Stage | Resource Types
.2+<.^| maxPerStageDescriptorSamplers
| sampler | combined image sampler
.3+<.^| maxPerStageDescriptorSampledImages
| sampled image | combined image sampler | uniform texel buffer
.2+<.^| maxPerStageDescriptorStorageImages
| storage image | storage texel buffer
.2+<.^| maxPerStageDescriptorUniformBuffers
| uniform buffer | uniform buffer dynamic
.2+<.^| maxPerStageDescriptorStorageBuffers
| storage buffer | storage buffer dynamic
| maxPerStageDescriptorInputAttachments
| input attachment^1^
|=============================
1::
Input attachments can: only be used in the fragment shader stage
[[interfaces-resources-layout]]
=== Offset and Stride Assignment
All variables with a storage class of code:PushConstant or code:Uniform must
be explicitly laid out using the code:Offset, code:ArrayStride, and
code:MatrixStride decorations. There are two different layouts requirements
depending on the specific resources.
[[interfaces-resources-layout-std140]]
*Standard Uniform Buffer Layout*
Member variables of an code:OpTypeStruct with storage class of
code:Uniform and a decoration of code:Block (uniform buffers) must: be laid
out according to the following rules.
* The code:Offset Decoration must be a multiple of its base alignment,
computed recursively as follows:
+
** a scalar of size latexmath:[$N$] has a base alignment of
latexmath:[$N$]
** a two-component vector, with components of size latexmath:[$N$], has
a base alignment of latexmath:[$2N$]
** a three- or four-component vector, with components of size
latexmath:[$N$], has a base alignment of latexmath:[$4N$]
** an array has a base alignment equal to the base alignment of its
element type, rounded up to a multiple of latexmath:[$16$]
** a structure has a base alignment equal to the largest base alignment
of any of its members, rounded up to a multiple of latexmath:[$16$]
** a row-major matrix of latexmath:[$C$] columns has a base alignment
equal to the base alignment of vector of latexmath:[$C$] matrix
components
** a column-major matrix has a base alignment equal to the base
alignment of the matrix column type
+
* Any code:ArrayStride or code:MatrixStride decoration must equal the base
alignment of the array or matrix from above.
[NOTE]
.Note
====
The *std140 layout* in GLSL satisfies these rules.
====
[[interfaces-resources-layout-std430]]
*Standard Storage Buffer Layout*
Member variables of an code:OpTypeStruct with a storage class of
code:PushConstant (push constants), or a storage class of code:Uniform
with a decoration of code:BufferBlock (storage buffers) must: be laid
out as <<interfaces-resources-layout-std140,above>>, except
for array and structure base alignment which do not need to be
rounded up to a multiple of latexmath:[$16$].
[NOTE]
.Note
====
The *std430 layout* in GLSL satisfies these rules.
====
[[interfaces-builtin-variables]]
== Built-In Variables
Built-in variables are accessed in shaders by declaring a variable decorated
using a code:BuiltIn decoration. The meaning of each code:BuiltIn decoration
is as follows. In the remainder of this section, the name of a built-in is
used interchangeably with a term equivalent to a variable decorated with
that particular built-in. Built-ins that represent integer values can: be
declared as either signed or unsigned 32-bit integers.
code:ClipDistance::
Variables decorated with the code:ClipDistance decoration provide the
mechanism for controlling user clipping. Declared as an array, the i^th^
element of the variable decorated as code:ClipDistance specifies a clip
distance for plane i. A clip distance of 0 means the vertex is on the plane,
a positive distance means the vertex is inside the clip half-space, and a
negative distance means the point is outside the clip half-space.
+
The code:ClipDistance array is explicitly sized by the shader.
+
The code:ClipDistance decoration can: be applied to array inputs in
tessellation control, tessellation evaluation and geometry shader stages
which will contain the values written by the previous stage. It can: be
applied to outputs in vertex, tessellation evaluation and geometry shaders.
In the last vertex processing stage, these values will be linearly
interpolated across the primitive and the portion of the primitive with
interpolated distances less than 0 will be considered outside the clip
volume.
+
In the fragment shader, the code:ClipDistance decoration can: be applied to
an array of floating-point input variables and contains the linearly
interpolated values described above.
+
code:ClipDistance mustnot: be used in compute shaders.
+
code:ClipDistance must: be declared as an array of 32-bit floating-point
values.
code:CullDistance::
A variable decorated as code:CullDistance provides a mechanism for a vertex
processing stage to reject an entire primitive. code:CullDistance can: be
applied to an array variable. If any member of this array is assigned a
negative value for all vertices belonging to a primitive, then the primitive
is discarded before rasterization. code:CullDistance can: be applied to an
output variable in the last vertex processing stage (vertex, tessellation
evaluation or geometry shader).
+
If applied to an input variable, that variable will contain the value of the
corresponding output in the previous shader stage. code:CullDistance
mustnot: be applied to an input in the vertex shader or to an output in the
fragment shader, and mustnot: be used in compute shaders.
+
In fragment shaders, the values of the code:CullDistance array are linearly
interpolated across each primitive.
+
code:CullDistance must: be declared as an array of 32-bit floating-point
values.
code:FragCoord::
This variable contains the framebuffer coordinate
latexmath:[$(x,y,z,\frac{1}{w})$] of the fragment being processed. The (x,y)
coordinate (0,0) is the upper left corner of the upper left pixel in the
framebuffer. The values of the x and y components of code:FragCoord reflect
the location of the center of the pixel (i.e. fractional values of
latexmath:[$(0.5,0.5)$]) when sample shading is not enabled, and the
location of the sample corresponding to the shader invocation when using
sample shading.
+
The z component of code:FragCoord is the interpolated depth value of the
primitive, and the w component is the interpolated
latexmath:[$\frac{1}{w}$].
+
The code:FragCoord decoration is only supported in fragment shaders. The
code:Centroid interpolation decoration is ignored on code:FragCoord.
+
code:FragCoord must: be declared as a four-component vector of 32-bit
floating-point values.
code:FragDepth::
Writing to an output variable decorated with code:FragDepth from the
fragment shader establishes a new depth value for all samples covered by the
fragment. This value will be used for depth testing and, if the depth test
passes, any subsequent write to the depth/stencil attachment. To write to
code:FragDepth, a shader must: declare the code:DepthReplacing execution
mode. If a shader declares the code:DepthReplacing execution mode and there
is an execution path through the shader that does not set code:FragDepth,
then the fragment's depth value is undefined for executions of the shader
that take that path. That is, if the fragment shader enables depth
replacing, then it must: always write it.
+
The code:FragDepth decoration is only supported in fragment shaders.
+
code:FragDepth must: be declared as a scalar 32-bit floating-point value.
code:FrontFacing::
The code:FrontFacing decoration can: be applied to an input variable in the
fragment shader. The value of this variable is non-zero if the current
fragment is considered to be part of a
<<primsrast-polygons-basic,front-facing>> primitive and is zero if the
fragment is considered to be part of a back-facing primitive.
+
--
The code:FrontFacing decoration is not available to shader stages other than
fragment.
code:FrontFacing must: be declared as a scalar 32-bit integer.
[NOTE]
.Note
====
In GLSL, code:gl_FrontFacing is declared as a code:bool. To achieve
similar semantics in SPIR-V, a variable of code:OpTypeBool can: be declared
and initialized as the result of the code:OpINotEqual operation with the
operands of the code:FrontFacing variable and an appropriately typed
constant zero.
====
--
code:GlobalInvocationID::
An input variable decorated with code:GlobalInvocationID will contain the
location of the current compute shader invocation within the global
workgroup. The value in this variable is equal to the index of the local
workgroup multiplied by the size of the local workgroup plus the value of
code:LocalInvocationID.
+
The code:GlobalInvocationID decoration is only supported in compute shaders.
+
code:GlobalInvocationID must: be declared as a three-component vector of
32-bit integers.
code:HelperInvocation::
This variable is non-zero if the fragment being shaded is a helper
invocation and zero otherwise. A helper invocation is an invocation of
the shader that is produced to satisfy internal requirements such as the
generation of derivatives.
+
--
The code:HelperInvocation decoration is only supported in fragment shaders.
code:HelperInvocation must: be declared as a scalar 32-bit integer.
[NOTE]
.Note
====
It is very likely that a helper invocation will have a value of
code:SampleMask fragment shader input value that is zero.
====
[NOTE]
.Note
====
In GLSL, code:HelperInvocation is declared as a code:bool. To achieve
similar semantics in SPIR-V, a variable of code:OpTypeBool can: be declared
and initialized as the result of the code:OpINotEqual operation with the
operands of the code:HelperInvocation variable and an appropriately typed
constant zero.
====
--
code:InvocationID::
In a geometry shader, an input variable decorated with the code:InvocationID
decoration contains the index of the current shader invocation, which ranges
from zero to the number of <<geometry-invocations,instances>> declared
in the shader. If the instance count of the geometry shader is one or is not
specified, then code:InvocationID will be zero.
+
In tessellation control shaders, and input variable decorated with the
code:InvocationID decoration contains the index of the output patch vertex
assigned to the tessellation control shader invocation.
+
The code:InvocationID decoration mustnot: be used in vertex, tessellation
evaluation, fragment, or compute shaders.
+
code:InvocationID must: be declared as a scalar 32-bit integer.
code:InstanceIndex::
The code:InstanceIndex decoration can: be applied to a vertex shader input
which will be filled with the index of the instance that is being processed
by the current vertex shader invocation. The value of code:InstanceIndex
begins at the value of the pname:firstInstance parameter to flink:vkCmdDraw
or flink:vkCmdDrawIndexed or at the value of the pname:firstInstance member
of a structure consumed by flink:vkCmdDrawIndirect or
flink:vkCmdDrawIndexedIndirect.
+
The code:InstanceIndex decoration mustnot: be used in any shader stage other
than vertex.
+
code:InstanceIndex must: be declared as a scalar 32-bit integer.
code:Layer::
The code:Layer decoration can: be applied to an output variable in the
geometry shader that is written with the framebuffer layer index to which
the primitive produced by the geometry shader will be directed. If a
geometry shader entry point's interface does not include an output variable
decorated with code:Layer, then the first layer is used. If a geometry
shader entry point's interface includes an output variable decorated with
code:Layer, it must: write the same value to code:Layer for all output
vertices of a given primitive. When used in a fragment shader, an input
variable decorated with code:Layer contains the layer index of the primitive
that the fragment invocation belongs to.
+
The code:Layer decoration is only supported in geometry and fragment
shaders.
+
code:Layer must: be declared as a scalar 32-bit integer.
code:LocalInvocationID::
This variable contains the location of the current compute shader invocation
within the local workgroup. The range of possible values for each component
of LocalInvocationID range from zero through the size of the workgroup (as
defined by code:LocalSize) in that dimension minus one. If the size of the
workgroup in a particular dimension is one, then the value of
LocalInvocationID in that dimension will be zero. That is, if the workgroup
is effectively two-dimensional, then pname:LocalInvocationID.z will be zero,
and if the workgroup is one-dimensional, then the values of both
pname:LocalInvocationID.y and pname:LocalInvocationID.z will be zero.
+
The code:LocalInvocationID decoration is only supported in compute shaders.
+
code:LocalInvocationID must: be declared as a three-component vector of
32-bit integers.
code:NumWorkGroups::
The code:NumWorkGroups decoration can: be applied to a code:uvec3 input
variable in a compute shader, in which case it will contain the number of
local workgroups that are part of the dispatch that the invocation belongs
to. It reflects the values passed to a call to flink:vkCmdDispatch or
through the structure consumed by the execution of
flink:vkCmdDispatchIndirect.
+
The code:NumWorkGroups decoration is only supported in compute shaders.
+
code:NumWorkGroups must: be declared as a three-component vector of 32-bit
integers.
code:PatchVertices::
An input variable decorated with code:PatchVertices in the tessellation
control or evaluation shader is an integer specifying the number of
vertices in the input patch being processed by the shader. A single
tessellation control or evaluation shader can: read patches of differing
sizes, so the value of the code:PatchVertices variable may: differ between
patches.
+
The code:PatchVertices decoration is only supported in tessellation control
and evaluation shaders.
+
code:PatchVertices must: be declared as scalar 32-bit integer.
code:PointCoord::
During point rasterization, a variable decorated with code:PointCoord
contains the coordinate of the current fragment within the point being
rasterized, normalized to the size of the point with origin in the upper
left corner of the point, as described in <<primsrast-points-basic,Basic
Point Rasterization>>. If the primitive the fragment shader invocation
belongs to is not a point then the value of code:PointCoord is undefined.
+
--
The code:PointCoord decoration is only supported in fragment shaders.
code:PointCoord must: be declared as two-component vector of 32-bit
floating-point values.
[NOTE]
.Note
====
Depending on how the point is rasterized, code:PointCoord may: never
reach (0,0) or (1,1).
====
--
code:PointSize::
The code:PointSize built-in decoration is used to pass the size of point
primitives between shader stages. It can: be applied to inputs to
tessellation control and geometry shaders. It can: be applied to output
variables in vertex, tessellation evaluation and geometry shaders. The value
written to the variable decorated as code:PointSize by the last vertex
processing stage in the pipeline is used as the framebuffer space size of
points produced by rasterization. As an input, it reflects the value written
to the output decorated with code:PointSize in the previous shader stage.
+
The code:PointSize decoration mustnot: be applied to inputs in the vertex
shader and mustnot: be used in fragment or compute shaders.
+
code:PointSize must: be declared as a scalar 32-bit floating-point value.
code:Position::
The code:Position built-in decoration can: be used on variables declared as
input to tessellation control, tessellation evaluation and geometry shaders.
It can: be used on variables declared as outputs in the vertex, tessellation
control, tessellation evaluation and geometry shaders. As an input, it
contains the data written to the output variable decorated as code:Position
in the previous shader stage. As an output, the data written to a variable
decorated as code:Position is passed to the next shader stage. In the last
vertex processing stage, the output position is used in subsequent primitive
assembly, clipping and rasterization operations.
+
Variables decorated as code:Position mustnot: be used as inputs in vertex
shaders and mustnot: be used in fragment or compute shaders.
+
code:Position must: be declared as a four-component vector of 32-bit
floating-point values.
code:PrimitiveID::
When the code:PrimitiveID decoration is applied to an input variable in the
tessellation control or tessellation evaluation shader, it will be filled
with the index of the patch within the current set of rendering primitives
that corresponds to the shader invocation.
+
When the code:PrimitiveID decoration is applied to an input variable in the
geometry shader, it will be filled with the number of primitives presented
as input to the geometry shader since the current set of rendering
primitives was started. When code:PrimitiveID is applied to an output in the
geometry shader, the resulting value is seen as an input to the fragment
shader.
+
When code:PrimitiveID is applied to an input in the fragment shader, it will
be filled with the primitive index written by the geometry shader if a
geometry shader is present, or with the value that would have been presented
as input to the geometry shader had it been present. If a geometry shader is
present and the fragment shader reads from an input variable decorated with
code:PrimitiveID, then the geometry shader must: write to an output variable
decorated with code:PrimitiveID in all execution paths, otherwise the value
of the code:PrimitiveID input in the fragment shader is undefined.
+
The code:PrimitiveID decoration mustnot: be used in vertex or compute
shaders. code:PrimitiveID mustnot: be used on output variables in
tessellation control, tessellation evaluation, or fragment shaders.
+
code:PrimitiveID must: be declared as scalar 32-bit integer.
code:SampleID::
The code:SampleID decoration can: be applied to an integer input variable in
the fragment shader. This variable will contain the zero-based index of the
sample the invocation corresponds to. The value of code:SampleID ranges from
zero to the number of samples in the framebuffer minus one. If a fragment
shader entry point's interface includes an input variable decorated with
code:SampleID, per-sample shading is enabled for draws that use that
fragment shader.
+
code:SampleID is not available in shader stages other than fragment.
+
code:SampleID must: be declared as a scalar 32-bit integer.
code:SampleMask::
A fragment input variable decorated with code:SampleMask will contain a
bitmask of the set of samples covered by the primitive generating the
fragment during rasterization. It has a sample bit set if and only if the
sample is considered covered for this fragment shader invocation.
code:SampleMask[] is an array of integers. Bits are mapped to samples in a
manner where bit B of mask M (`SampleMask[M]`) corresponds to sample
latexmath:[$32 \times M + B$].
+
When state specifies multiple fragment shader invocations for a given
fragment, the sample mask for any single fragment shader invocation
specifies the subset of the covered samples for the fragment that correspond
to the invocation. In this case, the bit corresponding to each covered
sample will be set in exactly one fragment shader invocation.
+
A fragment output variable decorated with code:SampleMask is an array of
integers forming a bit array in a manner similar an input variable decorated
with code:SampleMask, but where each bit represents coverage as computed by
the shader. Modifying the sample mask by writing zero to a bit of
code:SampleMask causes the sample to be considered uncovered. However,
setting sample mask bits to one will never enable samples not covered by the
original primitive. If the fragment shader is being evaluated at any
frequency other than per-fragment, bits of the sample mask not corresponding
to the current fragment shader invocation are ignored. This array must: be
sized in the fragment shader either implicitly or explicitly, to be no
larger than the implementation-dependent maximum sample-mask (as an array of
32-bit elements), determined by the maximum number of samples. If a fragment
shader entry point's interface includes an output variable decorated with
code:SampleMask, the sample mask will be undefined for any array elements of
any fragment shader invocations that fail to assign a value. If a fragment
shader entry point's interface does not include an output variable decorated
with code:SampleMask, the sample mask has no effect on the processing of a
fragment.
+
The code:SampleMask decoration is only supported in fragment shaders.
+
code:SampleMask must: be declared as an array of 32-bit integers.
code:SamplePosition::
This variable contains the sub-pixel position of the sample being shaded.
The top left of the pixel is considered to be at coordinate (0,0) and the
bottom right of the pixel is considered to be at coordinate (1,1). If a
fragment shader entry point's interface includes an input variable decorated
with code:SamplePosition, per-sample shading is enabled for draws that use
that fragment shader.
+
The code:SamplePosition decoration is only supported in fragment shaders.
+
code:SamplePosition must: be declared as a two-component vector of
floating-point values.
code:TessellationCoord::
The code:TessellationCoord is applied to an input variable in tessellation
evaluation shaders and specifies the three-dimensional (u,v,w) barycentric
coordinate of the tessellated vertex within the patch. The values of u, v,
and w are in the range latexmath:[$[0,1\]$] and vary linearly across the
primitive being subdivided. For the tessellation modes of code:Quads or
code:IsoLines, the third component is always zero.
+
The code:TessellationCoord decoration is only available to tessellation
evaluation shaders.
+
code:TessellationCoord must: be declared as three-component vector of 32-bit
floating-point values.
code:TessellationLevelOuter::
The code:TessellationLevelOuter decoration is used in tessellation control
shaders to decorate an output variable to contain the outer tessellation
factor for the resulting patch. This value is used by the tessellator
to control primitive tessellation and can: be read by
tessellation evaluation shaders. When applied to an input variable in a
tessellation evaluation shader, the shader can: read the value written by
the tessellation control shader.
+
The code:TessellationLevelOuter decoration is not available outside
tessellation control and evaluation shaders.
+
code:TessellationLevelOuter must: be declared as an array of size two,
containing 32-bit floating-point values.
code:TessellationLevelInner::
The code:TessellationLevelInner decoration is used in tessellation control
shaders to decorate an output variable to contain the inner tessellation
factor for the resulting patch. This value is used by the tessellator to
control primitive tessellation and can: be read by
tessellation evaluation shaders. When applied to an input variable in a
tessellation evaluation shader, the shader can: read the value written by
the tessellation control shader.
+
The code:TessellationLevelInner decoration is not available outside
tessellation control and evaluation shaders.
+
code:TessellationLevelInner must: be declared as an array of size four,
containing 32-bit floating-point values.
code:VertexIndex::
The code:VertexIndex decoration can: be applied to a vertex shader input
which will be filled with the index of the vertex that is being processed by
the current vertex shader invocation. For non-indexed draws, the value of
this variable begins at the value of the pname:firstVertex parameter to
flink:vkCmdDraw or the pname:firstVertex member of a structure consumed by
flink:vkCmdDrawIndirect and increments by one for each vertex in the draw.
For indexed draws, its value is the content of the index buffer for the
vertex plus the value of the pname:vertexOffset parameter to
flink:vkCmdDrawIndexed or the pname:vertexOffset member of the structure
consumed by flink:vkCmdDrawIndexedIndirect.
+
The value of code:VertexIndex starts at the same starting value for each
instance.
+
The code:VertexIndex decoration mustnot: be used in any shader stage other
than vertex.
+
code:VertexIndex must: be declared as a 32-bit integer.
code:ViewportIndex::
The code:ViewportIndex decoration can: be applied to an output variable in
the geometry shader that is written with the viewport index to which the
primitive produced by the geometry shader will be directed. The selected
viewport index is used to select the viewport transform and scissor
rectangle. If a geometry shader entry point's interface does not include an
output variable decorated with code:ViewportIndex, then the first viewport
is used. If a geometry shader entry point's interface includes an output
variable decorated with code:ViewportIndex, it must: write the same value to
code:ViewportIndex for all output vertices of a given primitive. When used
in a fragment shader, an input variable decorated with code:ViewportIndex
contains the viewport index of the primitive that the fragment invocation
belongs to.
+
The code:ViewportIndex decoration is only supported in geometry and fragment
shaders.
+
code:ViewportIndex must: be declared as a 32-bit integer.
code:WorkgroupID::
The code:WorkgroupID built-in decoration can: be applied to an input
variable in the compute shader. It will contain a three dimensional integer
index of the global workgroup that the current invocation is a member of.
Each component ranges from zero to the values of the parameters passed into
flink:vkCmdDispatch or read from the sname:VkDispatchIndirectCommand
structure read through a call to flink:vkCmdDispatchIndirect.
+
The code:WorkGroupID decoration is only supported in compute shaders.
+
code:WorkGroupID must: be declared as a three-component vector of 32-bit
integers.