Vulkan-Docs/doc/specs/vulkan/chapters/VK_AMD_shader_info.txt
Jon Leech 5769283448 Change log for January 15, 2018 Vulkan 1.0.68 spec update:
* Bump API patch number and header version number to 68 for this update.

Github Issues:

  * Added more details in the
    <<extended-functionality-extensions-compatibility, Extension
    Compatibility>> section, allowing explicit incompatibilities, and
    simplify corresponding language in the style guide, which now defers to
    the API Specification on this point (public issue 638).
  * Fix typo in description of slink:VkCommandBufferLevel::pname:level
    (public issue 651).
  * Only include extension-dependent valid usage statement for
    slink:VkImageSubresourceRange, and note that the extension names for
    header files described in the <<boilerplate-wsi-header, Window
    System-Specific Header Control>> section are only valid links, when the
    specification being viewed is built with the corresponding extensions
    enabled (public issue 652).

Internal Issues:

  * Add language to elink:VkResult specifying that when commands return an
    error, output parameter contents are undefined instead of unmodified
    (except for pname:sType and pname:pNext). Note that this is a behavior
    change. Add notes calling out slink:VkImageFormatProperties as an
    exception (internal issue 1118).
  * Add "`general-purpose`" to the style guide, and correct existing uses of
    "`general purpose`" as an adjective (internal issue 1121).
  * Add the ename:VK_DEBUG_REPORT_OBJECT_TYPE_VALIDATION_CACHE_EXT_EXT token
    for the `VK_EXT_validation_cache` extension, following the same naming
    pattern as other tokens in the extension, but keep the old
    ename:VK_DEBUG_REPORT_OBJECT_TYPE_VALIDATION_CACHE_EXT token around for
    backwards compatibility (internal issue 1126).

Other Issues:

  * Specify that flink:vkCmdSetDiscardRectangleEXT does not affect copies or
    clears, matching existing language for the scissor rectangle test.
  * Move the <<boilerplate-sType, pname:sType>> definition from the
    boilerplate appendix to the Fundamentals chapter, putting it together
    with the valid usage of pname:sType rather than having the definition
    split across two places.
  * Inline all of the etext:Vk*Flags definitions, moving each one from the
    boilerplate appendix to appear either after the corresponding
    etext:Vk*FlagBits value if one is defined, or after the first structure
    that includes them if not.
2018-01-15 05:29:25 -08:00

121 lines
5.2 KiB
Plaintext

// This section is included inside the Pipelines chapter (pipelines.txt)
[[pipelines-shader-information]]
== Pipeline Shader Information
[open,refpage='vkGetShaderInfoAMD',desc='Get information about a shader in a pipeline',type='protos']
--
Information about a particular shader that has been compiled as part of a
pipeline object can be extracted by calling:
include::../api/protos/vkGetShaderInfoAMD.txt[]
* pname:device is the device that created pname:pipeline.
* pname:pipeline is the target of the query.
* pname:shaderStage identifies the particular shader within the pipeline
about which information is being queried.
* pname:infoType describes what kind of information is being queried.
* pname:pInfoSize is a pointer to a value related to the amount of data
the query returns, as described below.
* pname:pInfo is either NULL or a pointer to a buffer.
If pname:pInfo is `NULL`, then the maximum size of the information that can:
be retrieved about the shader, in bytes, is returned in pname:pInfoSize.
Otherwise, pname:pInfoSize must: point to a variable set by the user to the
size of the buffer, in bytes, pointed to by pname:pInfo, and on return the
variable is overwritten with the amount of data actually written to
pname:pInfo.
If pname:pInfoSize is less than the maximum size that can: be retrieved by
the pipeline cache, then at most pname:pInfoSize bytes will be written to
pname:pInfo, and fname:vkGetShaderInfoAMD will return ename:VK_INCOMPLETE.
Not all information is available for every shader and implementations may
not support all kinds of information for any shader.
When a certain type of information is unavailable, the function returns
ename:VK_ERROR_FEATURE_NOT_PRESENT.
If information is successfully and fully queried, the function will return
ename:VK_SUCCESS.
For ename:VK_SHADER_INFO_TYPE_STATISTICS_AMD, an instance of
sname:VkShaderStatisticsInfoAMD will be written to the buffer pointed to by
pname:pInfo.
This structure will be populated with statistics regarding the physical
device resources used by that shader along with other miscellaneous
information and is described in further detail below.
For ename:VK_SHADER_INFO_TYPE_DISASSEMBLY_AMD, pname:pInfo points to a UTF-8
null-terminated string containing human-readable disassembly.
The exact formatting and contents of the disassembly string are
vendor-specific.
The formatting and contents of all other types of information, including
ename:VK_SHADER_INFO_TYPE_BINARY_AMD, are left to the vendor and are not
further specified by this extension.
include::../validity/protos/vkGetShaderInfoAMD.txt[]
--
[open,refpage='VkShaderStatisticsInfoAMD',desc='Statistical information about a particular shader within a pipeline',type='structs']
--
The sname:VkShaderStatisticsInfoAMD structure is defined as:
include::../api/structs/VkShaderStatisticsInfoAMD.txt[]
* pname:shaderStageMask are the combination of logical shader stages
contained within this shader.
* pname:resourceUsage is an instance of slink:VkShaderResourceUsageAMD
describing internal physical device resources used by this shader.
* pname:numPhysicalVgprs is the maximum number of vector instruction
general-purpose registers (VGPRs) available to the physical device.
* pname:numPhysicalSgprs is the maximum number of scalar instruction
general-purpose registers (SGPRs) available to the physical device.
* pname:numAvailableVgprs is the maximum limit of VGPRs made available to
the shader compiler.
* pname:numAvailableSgprs is the maximum limit of SGPRs made available to
the shader compiler.
* pname:computeWorkGroupSize is the local workgroup size of this shader in
{ X, Y, Z } dimensions.
Some implementations may merge multiple logical shader stages together in a
single shader.
In such cases, pname:shaderStageMask will contain a bitmask of all of the
stages that are active within that shader.
Consequently, if specifying those stages as input to
flink:vkGetShaderInfoAMD, the same output information may: be returned for
all such shader stage queries.
The number of available VGPRs and SGPRs (pname:numAvailableVgprs and
pname:numAvailableSgprs respectively) are the shader-addressable subset of
physical registers that is given as a limit to the compiler for register
assignment.
These values may: further be limited by implementations due to performance
optimizations where register pressure is a bottleneck.
include::../validity/structs/VkShaderStatisticsInfoAMD.txt[]
--
[open,refpage='VkShaderResourceUsageAMD',desc='Resource usage information about a particular shader within a pipeline',type='structs']
--
The sname:VkShaderResourceUsageAMD structure is defined as:
include::../api/structs/VkShaderResourceUsageAMD.txt[]
* pname:numUsedVgprs is the number of vector instruction general-purpose
registers used by this shader.
* pname:numUsedSgprs is the number of scalar instruction general-purpose
registers used by this shader.
* pname:ldsSizePerLocalWorkGroup is the maximum local data store size per
work group in bytes.
* pname:ldsUsageSizeInBytes is the LDS usage size in bytes per work group
by this shader.
* pname:scratchMemUsageInBytes is the scratch memory usage in bytes by
this shader.
include::../validity/structs/VkShaderResourceUsageAMD.txt[]
--