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

857 lines
41 KiB
Plaintext

// Copyright (c) 2015-2016 The Khronos Group Inc.
// Copyright notice at https://www.khronos.org/registry/speccopyright.html
[[renderpass]]
= Render Pass
A _render pass_ represents a collection of attachments, subpasses, and
dependencies between the subpasses, and describes how the attachments
are used over the course of the subpasses. The use of a render pass
in a command buffer is a _render pass instance_.
An _attachment description_ describes the properties of an attachment
including its format, sample count, and how its contents are
treated at the beginning and end of each render pass instance.
A _subpass_ represents a phase of rendering that reads and writes a
subset of the attachments in a render pass. Rendering commands are
recorded into a particular subpass of a render pass instance.
A _subpass description_ describes the subset of attachments that is involved
in the execution of a subpass. Each subpass can: read from some attachments
as _input attachments_, write to some as _color attachments_ or
_depth/stencil attachments_, and do resolve operations to others as _resolve
attachments_. A subpass description can: also include a set of _preserve
attachments_, which are attachments that are not read or written by the
subpass but whose contents must: be preserved throughout the subpass.
A subpass _uses_ an attachment if the attachment is a color, depth/stencil,
resolve, or input attachment for that subpass. A subpass does not use an
attachment if that attachment is preserved by the subpass. The first use of
an attachment is in the lowest numbered subpass that uses that attachment.
Similarly, the last use of an attachment is in the highest numbered subpass
that uses that attachment.
The subpasses in a render pass all render to the same dimensions, and
fragments for pixel (x,y,layer) in one subpass can: only read attachment
contents written by previous subpasses at that same (x,y,layer) location.
[NOTE]
.Note
====
By describing a complete set of subpasses a priori, render passes provide
the implementation an opportunity to optimize the storage and transfer of
attachment data between subpasses.
In practice, this means that subpasses with a simple framebuffer-space
dependency may: be merged into a single tiled rendering pass, keeping the
attachment data on-chip for the duration of a render pass instance. However,
it is also quite common for a render pass to only contain a single subpass.
====
_Subpass dependencies_ describe ordering restrictions between pairs of
subpasses. If no dependencies are specified, implementations may: reorder or
overlap portions (e.g., certain shader stages) of the execution of
subpasses. Dependencies limit the extent of overlap or reordering, and are
defined using masks of pipeline stages and memory access types. Each
dependency acts as an
<<synchronization-execution-and-memory-dependencies,execution and memory
dependency>>, similarly to how <<synchronization-pipeline-barriers,pipeline
barriers>> are defined. Dependencies are needed if two subpasses operate on
attachments with overlapping ranges of the same sname:VkDeviceMemory object
and at least one subpass writes to that range.
A _subpass dependency chain_ is a sequence of subpass dependencies in a
render pass, where the source subpass of each subpass dependency (after the
first) equals the destination subpass of the previous dependency.
A render pass describes the structure of subpasses and attachments
independent of any specific image views for the attachments.
The specific image views that will be used for the attachments, and their
dimensions, are specified in sname:VkFramebuffer objects. Framebuffers
are created with respect to a specific render pass that the framebuffer
is compatible with (see <<renderpass-compatibility,Render Pass
Compatibility>>). Collectively, a render pass and a framebuffer define the
complete render target state for one or more subpasses as well as the
algorithmic dependencies between the subpasses.
The various pipeline stages of the drawing commands for a given subpass may:
execute concurrently and/or out of order, both within and across drawing
commands. However for a given (x,y,layer,sample) sample location, certain
per-sample operations are performed in
<<fundamentals-queueoperation-apiorder,API order>>.
[[renderpass-creation]]
== Render Pass Creation
A render pass is created by calling:
include::../protos/vkCreateRenderPass.txt[]
* pname:device is the logical device that creates the render pass.
* pname:pCreateInfo is a pointer to an instance of the
sname:VkRenderPassCreateInfo structure that describes the parameters of
the render pass.
* pname:pAllocator controls host memory allocation as described in the
<<memory-allocation, Memory Allocation>> chapter.
* pname:pRenderPass points to a sname:VkRenderPass handle in which the
resulting render pass object is returned.
include::../validity/protos/vkCreateRenderPass.txt[]
The sname:VkRenderPassCreateInfo structure is defined as:
include::../structs/VkRenderPassCreateInfo.txt[]
* pname:sType is the type of this structure.
* pname:pNext is `NULL` or a pointer to an extension-specific structure.
* pname:flags is reserved for future use.
* pname:attachmentCount is the number of attachments used by this
render pass, or zero indicating no attachments. Attachments are referred
to by zero-based indices in the range [0,pname:attachmentCount).
* pname:pAttachments points to an array of pname:attachmentCount number of
slink:VkAttachmentDescription structures describing properties of the
attachments, or `NULL` if pname:attachmentCount is zero.
* pname:subpassCount is the number of subpasses to create for this render
pass. Subpasses are referred to by zero-based indices in the range
[0,pname:subpassCount). A render pass must: have at least one subpass.
* pname:pSubpasses points to an array of pname:subpassCount number of
slink:VkSubpassDescription structures describing properties of the
subpasses.
* pname:dependencyCount is the number of dependencies between pairs of
subpasses, or zero indicating no dependencies.
* pname:pDependencies points to an array of pname:dependencyCount number
of slink:VkSubpassDependency structures describing dependencies
between pairs of subpasses, or `NULL` if pname:dependencyCount is zero.
include::../validity/structs/VkRenderPassCreateInfo.txt[]
sname:VkAttachmentDescription is defined as:
include::../structs/VkAttachmentDescription.txt[]
* pname:format is a elink:VkFormat value specifying the format of the
image that will be used for the attachment.
* pname:samples is the number of samples of the image as defined
in elink:VkSampleCountFlagBits.
* pname:loadOp specifies how the contents of color and depth components of
the attachment are treated at the beginning of the subpass where it is
first used:
+
--
include::../enums/VkAttachmentLoadOp.txt[]
** ename:VK_ATTACHMENT_LOAD_OP_LOAD means the contents within the render
area will be preserved.
** ename:VK_ATTACHMENT_LOAD_OP_CLEAR means the contents within the render
area will be cleared to a uniform value, which is specified when a render
pass instance is begun.
** ename:VK_ATTACHMENT_LOAD_OP_DONT_CARE means the contents within the area
need not be preserved; the contents of the attachment will be undefined
inside the render area.
--
* pname:storeOp specifies how the contents of color and depth components
of the attachment are treated at the end of the subpass where it is last
used:
+
--
include::../enums/VkAttachmentStoreOp.txt[]
** ename:VK_ATTACHMENT_STORE_OP_STORE means the contents within the render
area are written to memory and will be available for reading after the
render pass instance completes once the writes have been synchronized
with ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT (for color attachments)
or ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT (for depth/stencil
attachments).
** ename:VK_ATTACHMENT_STORE_OP_DONT_CARE means the contents within the
render area are not needed after rendering, and may: be discarded; the
contents of the attachment will be undefined inside the render area.
--
* pname:stencilLoadOp specifies how the contents of stencil components of
the attachment are treated at the beginning of the subpass where it
is first used, and must: be one of the same values allowed for
pname:loadOp above.
* pname:stencilStoreOp specifies how the contents of stencil components of
the attachment are treated at the end of the last subpass where it
is used, and must: be one of the same values allowed for pname:storeOp
above.
* pname:initialLayout is the layout the attachment image subresource will
be in when a render pass instance begins.
* pname:finalLayout is the layout the attachment image subresource will be
transitioned to when a render pass instance ends. During a render pass
instance, an attachment can: use a different layout in each subpass, if
desired.
* pname:flags is a bitfield of elink:VkAttachmentDescriptionFlagBits
describing additional properties of the attachment:
include::../enums/VkAttachmentDescriptionFlagBits.txt[]
include::../validity/structs/VkAttachmentDescription.txt[]
If the attachment uses a color format, then pname:loadOp and pname:storeOp
are used, and pname:stencilLoadOp and pname:stencilStoreOp are ignored. If
the format has depth and/or stencil components, pname:loadOp and
pname:storeOp apply only to the depth data, while pname:stencilLoadOp and
pname:stencilStoreOp define how the stencil data is handled.
[[renderpass-precision]]
During a renderpass instance, input/color attachments with color formats
that have a component size of 8, 16, or 32 bits must: be represented in the
attachment's format throughout the instance. Attachments with other
floating- or fixed-point color formats, or with depth components may: be
represented in a format with a precision higher than the attachment format,
but must: be represented with the same range. When such a component is
loaded via the pname:loadOp, it will be converted into an
implementation-dependent format used by the render pass. Such components
must: be converted from the render pass format, to the format of the
attachment, before they are stored or resolved at the end of a render pass
instance via pname:storeOp. Conversions occur as described in
<<fundamentals-numerics,Numeric Representation and Computation>> and
<<fundamentals-fixedconv, Fixed-Point Data Conversions>>.
[[renderpass-aliasing]]
If pname:flags includes ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT, then
the attachment is treated as if it shares physical memory with another
attachment in the same render pass. This information limits the ability of
the implementation to reorder certain operations (like layout transitions
and the pname:loadOp) such that it is not improperly reordered against
other uses of the same physical memory via a different attachment. This is
described in more detail below.
If a render pass uses multiple attachments that alias the same device
memory, those attachments must: each include the
ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT bit in their attachment
description flags. Attachments aliasing the same memory occurs in
multiple ways:
* Multiple attachments being assigned the same image view as part of
framebuffer creation.
* Attachments using distinct image views that correspond to the same
subresource of an image.
* Attachments using views of distinct image subresources which are bound
to overlapping memory.
Render passes must: include subpass dependencies (either directly or via
a subpass dependency chain) between any two subpasses that operate on the
same attachment or aliasing attachments and those subpass dependencies must:
include execution and memory dependencies separating uses of the aliases, if
at least one of those subpasses writes to one of the aliases. Those
dependencies mustnot: include the ename:VK_DEPENDENCY_BY_REGION_BIT if the
aliases are views of distinct image subresources which overlap in memory.
Multiple attachments that alias the same memory mustnot: be used in a single
subpass. A given attachment index mustnot: be used multiple times in a
single subpass, with one exception: two subpass attachments can: use the
same attachment index if at least one use is as an input attachment and
neither use is as a resolve or preserve attachment. In other words, the same
view can: be used simultaneously as an input and color or depth/stencil
attachment, but mustnot: be used as multiple color or depth/stencil
attachments nor as resolve or preserve attachments. This valid scenario is
described in more detail <<renderpass-feedbackloop, below>>.
If a set of attachments alias each other, then all except the first to be
used in the render pass must: use an pname:initialLayout of
ename:VK_IMAGE_LAYOUT_UNDEFINED, since the earlier uses of the other aliases
make their contents undefined. Once an alias has been used and a different
alias has been used after it, the first alias mustnot: be used in any later
subpasses. However, an application can: assign the same image view to
multiple aliasing attachment indices, which allows that image view to be
used multiple times even if other aliases are used in between. Once an
attachment needs the ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT bit,
there should: be no additional cost of introducing additional aliases, and
using these additional aliases may: allow more efficient clearing of the
attachments on multiple uses via ename:VK_ATTACHMENT_LOAD_OP_CLEAR.
[NOTE]
.Note
====
The exact set of attachment indices that alias with each other is not known
until a framebuffer is created using the render pass, so the above
conditions cannot: be validated at render pass creation time.
====
sname:VkSubpassDescription is defined as:
include::../structs/VkSubpassDescription.txt[]
* pname:flags is reserved for future use.
* pname:pipelineBindPoint is a elink:VkPipelineBindPoint value specifying
whether this is a compute or graphics subpass. Currently, only graphics
subpasses are supported.
* pname:inputAttachmentCount is the number of input attachments.
* pname:pInputAttachments is an array of slink:VkAttachmentReference
structures (defined below) that lists which of the render pass's
attachments can: be read in the shader during the subpass, and what
layout the attachment images will be in during the subpass. Each element
of the array corresponds to an input attachment unit number in the
shader, i.e. if the shader declares an input variable
`layout(input_attachment_index=X, set=Y, binding=Z)` then it uses the
attachment provided in pname:pInputAttachments[X]. Input attachments
must: also be bound to the pipeline with a descriptor set, with the
input attachment descriptor written in the location (set=Y, binding=Z).
* pname:colorAttachmentCount is the number of color attachments.
* pname:pColorAttachments is an array of pname:colorAttachmentCount
slink:VkAttachmentReference structures that lists which of the render
pass's attachments will be used as color attachments in the subpass, and
what layout the attachment images will be in during the subpass. Each
element of the array correponds to a fragment shader output location,
i.e. if the shader declared an output variable `layout(location=X)` then
it uses the attachment provided in pname:pColorAttachments[X].
* pname:pResolveAttachments is `NULL` or a pointer to an array of
sname:VkAttachmentReference structures. If pname:pResolveAttachments is
not `NULL`, each of its elements corresponds to a color attachment (the
element in pname:pColorAttachments at the same index). At the end of
each subpass, the subpass's color attachments are resolved to
corresponding resolve attachments, unless the resolve attachment index
is ename:VK_ATTACHMENT_UNUSED or pname:pResolveAttachments is `NULL`. If
the first use of an attachment in a render pass is as a resolve
attachment, then the pname:loadOp is effectively ignored as the resolve
is guaranteed to overwrite all pixels in the render area.
* pname:pDepthStencilAttachment is a pointer to a
slink:VkAttachmentReference specifying which attachment will be used for
depth/stencil data and the layout it will be in during the subpass.
Setting the attachment index to ename:VK_ATTACHMENT_UNUSED or leaving
this pointer as `NULL` indicates that no depth/stencil attachment will
be used in the subpass.
* pname:preserveAttachmentCount is the number of preserved attachments.
* pname:pPreserveAttachments is an array of pname:preserveAttachmentCount
render pass attachment indices describing the attachments that
are not used by a subpass, but whose contents must: be preserved
throughout the subpass.
The contents of an attachment within the render area become undefined at
the start of a subpass S if all of the following conditions are true:
* The attachment is used as a color, depth/stencil, or resolve attachment
in any subpass in the render pass.
* There is a subpass S1 that uses or preserves the attachment, and a
subpass dependency from S1 to S.
* The attachment is not used or preserved in subpass S.
Once the contents of an attachment become undefined in subpass S, they
remain undefined for subpasses in subpass dependency chains starting with
subpass S until they are written again. However, they remain valid for
subpasses in other subpass dependency chains starting with subpass S1 if
those subpasses use or preserve the attachment.
include::../validity/structs/VkSubpassDescription.txt[]
The sname:VkAttachmentReference structure is defined as:
include::../structs/VkAttachmentReference.txt[]
* pname:attachment is the index of the attachment of the render pass, and
corresponds to the index of the corresponding element in the
pname:pAttachments array of the sname:VkRenderPassCreateInfo structure.
If any color or depth/stencil attachments are
ename:VK_ATTACHMENT_UNUSED, then no writes occur for those attachments.
* pname:layout is a elink:VkImageLayout value specifying the layout the
attachment uses during the subpass. The implementation will
automatically perform layout transitions as needed between subpasses to
make each subpass use the requested layouts.
include::../validity/structs/VkAttachmentReference.txt[]
The sname:VkSubpassDependency structure is defined as:
include::../structs/VkSubpassDependency.txt[]
* pname:srcSubpass and pname:dstSubpass are the subpass indexes of the
producer and consumer subpasses, respectively. pname:srcSubpass and
pname:dstSubpass can: also have the special value
ename:VK_SUBPASS_EXTERNAL. The source subpass must: always be a lower
numbered subpass than the destination subpass (excluding external
subpasses and
<<synchronization-pipeline-barriers-subpass-self-dependencies,
self-dependencies>>), so that the order of subpass descriptions is a
valid execution ordering, avoiding cycles in the dependency graph.
* pname:srcStageMask, pname:dstStageMask, pname:srcAccessMask,
pname:dstAccessMask, and pname:dependencyFlags describe an
<<synchronization-execution-and-memory-dependencies,execution and memory
dependency>> between subpasses. The bits that can: be included in
pname:dependencyFlags are:
+
include::../enums/VkDependencyFlagBits.txt[]
** If pname:dependencyFlags contains ename:VK_DEPENDENCY_BY_REGION_BIT,
then the dependency is by-region as defined in
<<synchronization-execution-and-memory-dependencies,Execution And Memory
Dependencies>>.
Each subpass dependency defines an execution and memory dependency
between two sets of commands, with the second set depending on the first
set. When pname:srcSubpass does not equal pname:dstSubpass then the first
set of commands is:
* All commands in the subpass indicated by pname:srcSubpass, if
pname:srcSubpass is not ename:VK_SUBPASS_EXTERNAL.
* All commands before the render pass instance, if pname:srcSubpass is
ename:VK_SUBPASS_EXTERNAL.
While the corresponding second set of commands is:
* All commands in the subpass indicated by pname:dstSubpass, if
pname:dstSubpass is not ename:VK_SUBPASS_EXTERNAL.
* All commands after the render pass instance, if pname:dstSubpass is
ename:VK_SUBPASS_EXTERNAL.
When pname:srcSubpass equals pname:dstSubpass then the first set consists of
commands in the subpass before a call to flink:vkCmdPipelineBarrier and the
second set consists of commands in the subpass following that same call as
described in the
<<synchronization-pipeline-barriers-subpass-self-dependencies, Subpass
Self-dependency>> section.
The pname:srcStageMask, pname:dstStageMask, pname:srcAccessMask,
pname:dstAccessMask, and pname:dependencyFlags parameters of the dependency
are interpreted the same way as for other dependencies, as described in
<<synchronization, Synchronization and Cache Control>>.
include::../validity/structs/VkSubpassDependency.txt[]
Automatic image layout transitions between subpasses also interact with the
subpass dependencies. If two subpasses are connected by a dependency and
those two subpasses use the same attachment in a different layout, then the
layout transition will occur after the memory accesses via
pname:srcAccessMask have completed in all pipeline stages included in
pname:srcStageMask in the source subpass, and before any memory accesses
via pname:dstAccessMask occur in any pipeline stages included in
pname:dstStageMask in the destination subpass.
The automatic image layout transitions from pname:initialLayout to the first
used layout (if it is different) are performed according to the following
rules:
* If the attachment does not include the
ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT bit and there is no
subpass dependency from ename:VK_SUBPASS_EXTERNAL to the first subpass
that uses the attachment, then it is as if there were such a dependency
with pname:srcStageMask = pname:srcAccessMask = 0 and pname:dstStageMask
and pname:dstAccessMask including all relevant bits (all graphics
pipeline stages and all access types that use image resources), with the
transition executing as part of that dependency. In other words, it may:
overlap work before the render pass instance and is complete before the
subpass begins.
* If the attachment does not include the
ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT bit and there is a subpass
dependency from ename:VK_SUBPASS_EXTERNAL to the first subpass that uses
the attachment, then the transition executes as part of that dependency
and according to its stage and access masks. It mustnot: overlap work
that came before the render pass instance that is included in the source
masks, but it may: overlap work in previous subpasses.
* If the attachment includes the
ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT bit, then the transition
executes according to all the subpass dependencies with pname:dstSubpass
equal to the first subpass index that the attachment is used in. That
is, it occurs after all memory accesses in the source stages and masks
from all the source subpasses have completed and are available, and
before the union of all the destination stages begin, and the new layout
is visible to the union of all the destination access types. If there
are no incoming subpass dependencies, then this case follows the first
rule.
Similar rules apply for the transition to the pname:finalLayout, using
dependencies with pname:dstSubpass equal to ename:VK_SUBPASS_EXTERNAL
If an attachment specifies the ename:VK_ATTACHMENT_LOAD_OP_CLEAR load
operation, then it will logically be cleared at the start of the first
subpass where it is used.
[NOTE]
.Note
====
Implementations may: move clears earlier as long as it does not affect the
operation of a render pass instance. For example, an implementation may:
choose to clear all attachments at the start of the render pass instance. If
an attachment has the ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT flag
set, then the clear must: occur at the start of subpass where the attachment
is first used, in order to preserve the operation of the render pass
instance.
====
The first use of an attachment mustnot: specify a layout equal to
ename:VK_IMAGE_LAYOUT_DEPTH_STENCIL_READ_ONLY_OPTIMAL or
ename:VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL if the attachment specifies
that the pname:loadOp is ename:VK_ATTACHMENT_LOAD_OP_CLEAR. If a subpass
uses the same attachment as both an input attachment and either a color
attachment or a depth/stencil attachment, then both uses must: observe the
result of the clear.
Similarly, if an attachment specifies that the pname:storeOp is
ename:VK_ATTACHMENT_STORE_OP_STORE, then it will logically be stored at the
end of the last subpass where it is used.
[NOTE]
.Note
====
Implementations may: move stores
later as long as it does not affect the operation of a render pass instance.
If an attachment has the ename:VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT flag
set, then the store must: occur at the end of the highest numbered subpass
that uses the attachment.
====
If an attachment is not used by any subpass, then the pname:loadOp and the
pname:storeOp are ignored and the attachment's memory contents will not be
modified by execution of a render pass instance.
It will be common for a render pass to consist of a simple linear graph of
dependencies, where subpass N depends on subpass N-1 for all N, and the
operation of the memory barriers and layout transitions is fairly
straightforward to reason about for those simple cases. But for more complex
graphs, there are some rules that govern when there must be dependencies
between subpasses.
As stated earlier, render passes must: include subpass dependencies which
(either directly or via a subpass dependency chain) separate any two
subpasses that operate on the same attachment or aliasing attachments, if at
least one of those subpasses writes to the attachment. If an image layout
changes between those two subpasses, the implementation uses the stageMasks
and accessMasks indicated by the subpass dependency as the masks that
control when the layout transition must: occur. If there is not a layout
change on the attachment, or if an implementation treats the two layouts
identically, then it may: treat the dependency as a simple execution/memory
barrier.
If two subpasses use the same attachment in different layouts but both uses
are read-only (i.e. input attachment, or read-only depth/stencil
attachment), the application does not need to express a dependency between
the two subpasses. Implementations that treat the two layouts differently
may: deduce and insert a dependency between the subpasses, with the
implementation choosing the appropriate stage masks and access masks based
on whether the attachment is used as an input or depth/stencil attachment,
and may: insert the appropriate layout transition along with the
execution/memory barrier. Implementations that treat the two layouts
identically need not insert a barrier, and the two subpasses may: execute
simultaneously. The stage masks and access masks are chosen as follows:
- for input attachments, stage mask =
ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, access mask =
ename:VK_ACCESS_INPUT_ATTACHMENT_READ_BIT.
- for depth/stencil attachments, stage mask =
ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT |
ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT, access mask =
ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT
where pname:srcStageMask and pname:srcAccessMask are taken based on usage in
the source subpass and pname:dstStageMask and pname:dstAccessMask are taken
based on usage in the destination subpass.
[[renderpass-feedbackloop]]
If a subpass uses the same attachment as both an input attachment and either
a color attachment or a depth/stencil attachment, reads from the input
attachment are not automatically coherent with writes through the color or
depth/stencil attachment. In order to achieve well-defined results, one of
two criteria must: be satisfied. First, if the color components or
depth/stencil components read by the input attachment are mutually exclusive
with the components written by the color or depth/stencil
attachment then there is no _feedback loop_ and the reads and writes both
function normally, with the reads observing values from the previous
subpass(es) or from memory. This option requires the graphics pipelines
used by the subpass to disable writes to color components that are read as
inputs via the pname:colorWriteMask, and to disable writes to depth/stencil
components that are read as inputs via pname:depthWriteEnable or
pname:stencilTestEnable.
Second, if the input attachment reads components that are written by the
color or depth/stencil attachment, then there is a feedback loop and a
pipeline barrier must: be used between when the attachment is written and
when it is subsequently read by later fragments. This pipeline barrier must:
follow the rules of a self-dependency as described in
<<synchronization-pipeline-barriers-subpass-self-dependencies,Subpass
Self-dependency>>, where the barrier's flags include:
* pname:dstStageMask = ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT,
* pname:dstAccessMask = ename:VK_ACCESS_INPUT_ATTACHMENT_READ_BIT, and
* pname:srcAccessMask = ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT (for
color attachments) or pname:srcAccessMask =
ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT (for depth/stencil
attachments).
* pname:srcStageMask = ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT
(for color attachments) or pname:srcStageMask =
ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |
ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT (for depth/stencil
attachments).
* pname:dependencyFlags = ename:VK_DEPENDENCY_BY_REGION_BIT.
A pipeline barrier is needed each time a fragment will read a particular
(x,y,layer,sample) location if that location has been written since the most
recent pipeline barrier, or since the start of the subpass if there have
been no pipeline barriers since the start of the subpass.
An attachment used as both an input attachment and color attachment must: be
in the ename:VK_IMAGE_LAYOUT_GENERAL layout. An attachment used as both an
input attachment and depth/stencil attachment must: be in either the
ename:VK_IMAGE_LAYOUT_GENERAL or
ename:VK_IMAGE_LAYOUT_DEPTH_STENCIL_READ_ONLY_OPTIMAL layout. Since an
attachment in the ename:VK_IMAGE_LAYOUT_DEPTH_STENCIL_READ_ONLY_OPTIMAL
layout is read-only, this situation is not a feedback loop.
To destroy a render pass, call:
include::../protos/vkDestroyRenderPass.txt[]
* pname:device is the logical device that destroys the render pass.
* pname:renderPass is the handle of the render pass to destroy.
* pname:pAllocator controls host memory allocation as described in the
<<memory-allocation, Memory Allocation>> chapter.
include::../validity/protos/vkDestroyRenderPass.txt[]
[[renderpass-compatibility]]
== Render Pass Compatibility
Framebuffers and graphics pipelines are created based on a specific
render pass object. They must: only be used with that render pass object, or
one compatible with it.
Two attachment references are compatible if they have matching format and
sample count, or are both ename:VK_ATTACHMENT_UNUSED or the pointer that
would contain the reference is `NULL`.
Two arrays of attachment references are compatible if all corresponding
pairs of attachments are compatible. If the arrays are of different lengths,
attachment references not present in the smaller array are treated as
ename:VK_ATTACHMENT_UNUSED.
Two render passes that contain only a single subpass are compatible if
their corresponding color, input, resolve, and depth/stencil attachment
references are compatible.
If two render passes contain more than one subpass, they are compatible if
they are identical except for:
* Initial and final image layout in attachment descriptions
* Load and store operations in attachment descriptions
* Image layout in attachment references
A framebuffer is compatible with a render pass if it was created using the
same render pass or a compatible render pass.
== Framebuffers
Render passes operate in conjunction with framebuffers, which represent a
collection of specific memory attachments that a render pass instance uses.
An application creates a framebuffer by calling:
include::../protos/vkCreateFramebuffer.txt[]
* pname:device is the logical device that creates the framebuffer.
* pname:pCreateInfo points to a sname:VkFramebufferCreateInfo structure
which describes additional information about framebuffer creation.
* pname:pAllocator controls host memory allocation as described in the
<<memory-allocation, Memory Allocation>> chapter.
* pname:pFramebuffer points to a sname:VkFramebuffer handle in which the
resulting framebuffer object is returned.
include::../validity/protos/vkCreateFramebuffer.txt[]
The sname:VkFramebufferCreateInfo structure is defined as:
include::../structs/VkFramebufferCreateInfo.txt[]
* pname:sType is the type of this structure.
* pname:pNext is `NULL` or a pointer to an extension-specific structure.
* pname:flags is reserved for future use.
* pname:renderPass is a render pass that defines what render passes the
framebuffer will be compatible with. See
<<renderpass-compatibility,Render Pass Compatibility>> for details.
* pname:attachmentCount is the number of attachments.
* pname:pAttachments is an array of sname:VkImageView handles, each of
which will be used as the corresponding attachment in a render pass
instance.
* pname:width, pname:height and pname:layers define the dimensions of the
framebuffer.
include::../validity/structs/VkFramebufferCreateInfo.txt[]
Image subresources used as attachments mustnot: be used via any
non-attachment usage for the duration of a render pass instance.
[NOTE]
.Note
====
This restriction means that the render pass has full knowledge of all uses
of all of the attachments, so that the implementation is able to make
correct decisions about when and how to perform layout transitions, when to
overlap execution of subpasses, etc.
====
[[renderpass-noattachments]]
It is legal for a subpass to use no color or depth/stencil attachments, and
rather use shader side effects such as image stores and atomics to produce
an output. In this case, the subpass continues to use the pname:width,
pname:height, and pname:layers of the framebuffer to define the dimensions
of the rendering area, and the pname:rasterizationSamples from each
pipeline's sname:VkPipelineMultisampleStateCreateInfo to define the number
of samples used in rasterization; however, if
sname:VkPhysicalDeviceFeatures::pname:variableMultisampleRate is
code:VK_FALSE, then all pipelines to be bound with a given zero-attachment
subpass must: have the same value for
sname:VkPipelineMultisampleStateCreateInfo::pname:rasterizationSamples.
To destroy a framebuffer, call:
include::../protos/vkDestroyFramebuffer.txt[]
* pname:device is the logical device that destroys the framebuffer.
* pname:framebuffer is the handle of the framebuffer to destroy.
* pname:pAllocator controls host memory allocation as described in the
<<memory-allocation, Memory Allocation>> chapter.
include::../validity/protos/vkDestroyFramebuffer.txt[]
[[renderpass-commands]]
== Render Pass Commands
An application records the commands for a render pass instance one subpass
at a time, by beginning a render pass instance, iterating over the subpasses
to record commands for that subpass, and then ending the render pass
instance.
To begin a render pass instance, call:
include::../protos/vkCmdBeginRenderPass.txt[]
* pname:commandBuffer is the command buffer in which to record the
command.
* pname:pRenderPassBegin is a pointer to a sname:VkRenderPassBeginInfo
structure (defined below) which indicates the render pass to begin an
instance of, and the framebuffer the instance uses.
* pname:contents specifies how the commands in the first subpass will be
provided, and is one of the values:
+
--
include::../enums/VkSubpassContents.txt[]
If pname:contents is ename:VK_SUBPASS_CONTENTS_INLINE, the contents of the
subpass will be recorded inline in the primary command buffer, and secondary
command buffers mustnot: be executed within the subpass. If pname:contents
is ename:VK_SUBPASS_CONTENTS_SECONDARY_COMMAND_BUFFERS, the contents are
recorded in secondary command buffers that will be called from the primary
command buffer, and fname:vkCmdExecuteCommands is the only valid command on
the command buffer until fname:vkCmdNextSubpass or fname:vkCmdEndRenderPass.
--
include::../validity/protos/vkCmdBeginRenderPass.txt[]
After beginning a render pass instance, the command buffer is ready to
record the commands for the first subpass of that render pass.
[[renderpass-VkRenderPassBeginInfo]]
The sname:VkRenderPassBeginInfo structure is defined as:
include::../structs/VkRenderPassBeginInfo.txt[]
* pname:sType is the type of this structure.
* pname:pNext is `NULL` or a pointer to an extension-specific structure.
* pname:renderPass is the render pass to begin an instance of.
* pname:framebuffer is the framebuffer containing the attachments that are
used with the render pass.
* pname:renderArea is the render area that is affected by the render pass
instance, and is described in more detail below.
* pname:clearValueCount is the number of elements in pname:pClearValues.
* pname:pClearValues is an array of sname:VkClearValue structures that
contains clear values for each attachment, if the attachment uses a
pname:loadOp value of ename:VK_ATTACHMENT_LOAD_OP_CLEAR. The array is
indexed by attachment number, with elements corresponding to uncleared
attachments being unused.
include::../validity/structs/VkRenderPassBeginInfo.txt[]
pname:renderArea is the render area that is affected by the render pass
instance. The effects of attachment load, store and resolve operations are
restricted to the pixels whose x and y coordinates fall within the render
area on all attachments. The render area extends to all layers of
pname:framebuffer. The application must: ensure (using scissor if necessary)
that all rendering is contained within the render area, otherwise the pixels
outside of the render area become undefined and shader side effects may: or
maynot: occur for fragments outside the render area. The render area must:
be contained within the framebuffer dimensions.
[NOTE]
.Note
====
There may: be a performance cost for using a render area smaller than the
framebuffer, unless it matches the render area granularity for the render
pass.
====
The render area granularity is queried by calling:
include::../protos/vkGetRenderAreaGranularity.txt[]
* pname:device is the logical device that owns the render pass.
* pname:renderPass is a handle to a render pass.
* pname:pGranularity points to a sname:VkExtent2D structure in which the
granularity is returned.
include::../validity/protos/vkGetRenderAreaGranularity.txt[]
The conditions leading to an optimal pname:renderArea are:
* the pname:offset.x member in pname:renderArea is a multiple of the
pname:width member of the returned sname:VkExtent2D (the horizontal
granularity).
* the pname:offset.y member in pname:renderArea is a multiple of the
pname:height of the returned sname:VkExtent2D (the vertical
granularity).
* either the pname:offset.width member in pname:renderArea is a multiple
of the horizontal granularity or pname:offset.x+pname:offset.width is
equal to the pname:width of the pname:framebuffer in the
sname:VkRenderPassBeginInfo.
* either the pname:offset.height member in pname:renderArea is a multiple
of the vertical granularity or pname:offset.y+pname:offset.height is
equal to the pname:height of the pname:framebuffer in the
sname:VkRenderPassBeginInfo.
After recording the commands for a subpass, an application transitions to
the next subpass in the render pass instance by calling:
include::../protos/vkCmdNextSubpass.txt[]
* pname:commandBuffer is the command buffer in which to record the
command.
* pname:contents specifies how the commands in the next subpass will be
provided, in the same fashion as the corresponding parameter of
flink:vkCmdBeginRenderPass.
include::../validity/protos/vkCmdNextSubpass.txt[]
The subpasses indices for a render pass begin at zero when
fname:vkCmdBeginRenderPass is recorded, and increments each time
fname:vkCmdNextSubpass is recorded.
Moving to the next subpass automatically performs any multisample resolve
operations in the subpass being ended. End-of-subpass multisample resolves
are treated as color attachment writes for the purposes of synchronization.
That is, they are considered to execute in the
ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT pipeline stage and their
writes are synchronized with ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT.
Synchronization between rendering within a subpass and any resolve
operations at the end of the subpass occurs automatically, without need for
explicit dependencies or pipeline barriers. However, if the resolve
attachment is also used in a different subpass, an explicit dependency is
needed.
After transitioning to the next subpass, the application can: record the
commands for that subpass.
After recording the commands for the last subpass, an application records a
command to end a render pass instance by calling:
include::../protos/vkCmdEndRenderPass.txt[]
* pname:commandBuffer is the command buffer in which to end the current
render pass instance.
include::../validity/protos/vkCmdEndRenderPass.txt[]
Ending a render pass instance performs any multisample resolve operations on
the final subpass.