| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/chapters/VK_EXT_pipeline_creation_feedback/ |
| D | pipelines.adoc | 11 …nFeedbackCreateInfo',desc='Request for feedback about the creation of a pipeline',type='structs',x…
13 Feedback about the creation of a particular pipeline object can: be obtained
40 An implementation should: write pipeline creation feedback to
41 pname:pPipelineCreationFeedback and may: write pipeline stage creation
74 …kPipelineCreationFeedback',desc='Feedback about the creation of a pipeline or pipeline stage',type…
87 providing feedback about the creation of a pipeline or of a pipeline
89 * pname:duration is the duration spent creating a pipeline or pipeline
100 [open,refpage='VkPipelineCreationFeedbackFlagBits',desc='Bitmask specifying pipeline or pipeline st…
116 indicates that a readily usable pipeline or pipeline stage was found in
117 the pname:pipelineCache specified by the application in the pipeline
[all …]
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/appendices/ |
| D | VK_EXT_pipeline_properties.adoc | 22 In order to support this, the pipeline state is represented in a
27 application and use a layer to record and serialize the pipeline state and
29 Each pipeline is represented by a separate JSON file, and can be identified
32 Once the pipelines have been compiled by the offline pipeline cache
34 pname:pipelineIdentifier for identifying the pipeline via Vulkan SC's
38 pname:pipelineIdentifier associated with each pipeline so that the
39 application can store this with its pipeline metadata and the Vulkan SC
41 pipeline cache.
58 (2) This is intended to be a general pipeline properties query, but is
59 currently only retrieving the pipeline identifier.
[all …]
|
| D | VK_EXT_shader_module_identifier.adoc | 24 When pipeline caches are primed, either explicitly through e.g.
34 A pipeline creation call with such a module may: succeed if a pipeline could
44 pipeline is that pipeline identifiers change when a driver is updated, but
48 pipeline caches ahead of time.
49 When on-disk pipeline caches are updated, the same shader identifiers could
50 lead to a pipeline cache hit.
|
| D | VK_EXT_pipeline_creation_cache_control.adoc | 33 predictability of pipeline creation cost.
35 within the client driver during pipeline creation to the application before
43 calls for pipeline creation.
48 Proper usage of pipeline caches and/or derivative pipelines help mitigate
52 the workload of pipeline creation including possible shader compilation.
54 Applications making API calls to create a pipeline must be prepared for any
76 When set, an ICD must not attempt pipeline or shader compilation to create
77 the pipeline object.
78 In such a case, if the implementation fails to create a pipeline without
81 the pipeline.
[all …]
|
| D | VK_KHR_pipeline_library.adoc | 18 A pipeline library is a special pipeline that cannot be bound, instead it
21 This extension defines the infrastructure for pipeline libraries, but does
22 not specify the creation or usage of pipeline libraries.
|
| D | VK_EXT_pipeline_library_group_handles.adoc | 24 When using pipeline libraries in ray tracing pipelines, a library might get
26 An application can have a strategy where a ray tracing pipeline is comprised
27 of N pipeline libraries and is later augumented by creating a new pipeline
30 handle is tied to the pipeline, not the library.
35 directly from pipeline libraries.
|
| D | VK_KHR_pipeline_executable_properties.adoc | 30 When a pipeline is created, its state and shaders are compiled into zero or
32 against that pipeline.
34 different executables produced by the pipeline compilation process.
44 1) What should we call the pieces of the pipeline which are produced by the
49 determined that "`pipeline binary`" could have a fairly broad meaning (such
50 as a binary serialized form of an entire pipeline) and was too big of a
|
| D | VK_HUAWEI_subpass_shading.adoc | 24 This extension allows applications to execute a subpass shading pipeline in
27 A subpass shading pipeline is a pipeline with the compute pipeline ability,
36 The code:GlobalInvocationId.xy of a subpass shading pipeline is equal to the
37 code:FragCoord.xy of a graphic pipeline in the same render pass subtracted
47 This extension allows a subpass's pipeline bind point to be
213 Example of subpass shading pipeline creation
247 VKPipeline pipeline;
250 …putePipelines(device, pipelineCache, 1, &subpassShadingComputePipelineCreateInfo, NULL, &pipeline);
260 naming of other pipeline stages
|
| D | VK_EXT_subgroup_size_control.adoc | 45 physical device supports and a pipeline create flag to enable that pipeline
48 modules provided to pipeline creation may: vary between the
53 required subgroup size for a given pipeline stage.
55 support a required subgroup size>>, and any pipeline of a supported stage
57 structure to set the subgroup size for that shader stage of the pipeline.
67 Developers can also specify a new pipeline shader stage create flag that
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/chapters/ |
| D | pipelines.adoc | 12 handled by the various pipeline stages, or control data transfer between
14 Commands are effectively sent through a processing pipeline, either a
22 The graphics pipeline can be operated in two modes, as either _primitive
23 shading_ or _mesh shading_ pipeline.
29 The first stage of the <<pipelines-graphics,graphics pipeline>>
55 When using the <<mesh,_mesh shading_>> pipeline input primitives are not
58 The work on the mesh pipeline is initiated by the application
87 The <<pipelines-compute,compute pipeline>> is a separate pipeline from the
88 graphics pipeline, which operates on one-, two-, or three-dimensional
94 Actual ordering guarantees between pipeline stages are explained in detail
[all …]
|
| D | executiongraphs.adoc | 10 To achieve this, a new execution graph pipeline is provided, that links
13 Each linked pipeline or shader describes an _execution node_ within the
22 [open,refpage='vkCreateExecutionGraphPipelinesAMDX',desc='Creates a new execution graph pipeline ob…
31 pipeline caching is disabled; or the handle of a valid
32 <<pipelines-cache,pipeline cache>> object, in which case use of that
41 which the resulting execution graph pipeline objects are returned.
43 The implementation will create a pipeline in each element of
45 If creation of any pipeline fails, that pipeline will be set to
49 If creation fails for a pipeline create info with a
68 the ename:VK_PIPELINE_CREATE_DERIVATIVE_BIT flag, the base pipeline
[all …]
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/proposals/ |
| D | VK_EXT_pipeline_library_group_handles.adoc | 10 …erying ray tracing group handles on library creation time rather than waiting for pipeline linking.
15 AddToStateObject() is another API which incrementally adds shader groups to a ray tracing pipeline …
19 When using pipeline libraries with ray tracing,
20 the intent is that pipeline libraries allow for a flexible model where pipeline libraries are compi…
21 Based on the relevant materials that are active in a scene, a ray tracing pipeline can be linked th…
23 However, relinking a pipeline in the current API means that all group handles are invalidated,
24 and group handles must be required and uploaded to the device for every pipeline. This is not desir…
29 The only way to expose this feature is to allow group handles to be queried on the pipeline library…
31 For every library that includes a pipeline library,
35 …lso allow capture replay handles to be extracted from pipeline libraries. When creating a pipeline…
[all …]
|
| D | VK_EXT_pipeline_protected_access.adoc | 10 This proposal regards pipeline access to protected memory, and provides the
18 As this feature is enabled on the device, every pipeline created by the driver
21 pipeline creation and/or execution.
28 reduced performance with every pipeline instead of only those that do in fact
32 protected memory access in pipeline granularity.
37 specify whether and how each individual pipeline may access protected memory.
42 the pipeline from being used in a protected command buffer and submission.
45 of a pipeline to protected command buffers.
60 access when creating a pipeline; a set of `VkBufferUsageFlags` and
81 - This can theoretically lead to the most efficient pipeline that only pays a
[all …]
|
| D | VK_EXT_graphics_pipeline_library.adoc | 11 of pipelines, improving the performance of pipeline compilation for
26 pipeline cache is nearly unmanageable.
41 regress the intended wins from moving to pipeline objects – there should be
57 . Separately compiled pipeline/state blobs
103 applications to create partial pipelines (pipeline libraries) containing
105 These pipeline libraries can be linked together to form a final executable.
150 a full pipeline compilation.
170 the final linked pipeline.
193 For all pipeline libraries
200 libraries or for any final pipeline.
[all …]
|
| D | VK_EXT_dynamic_rendering_unused_attachments.adoc | 12 many more pipeline objects to be created to work around the limitation.
18 multiple variants of the pipeline object in order to be compatible with the render
20 These restrictions were put in place to accommodate hardware that uses the pipeline
22 change the framebuffer layout during the render pass instance, requiring all pipeline
36 This can cause a combinatorial explosion of pipeline objects, depending on the
37 application, or require pipeline objects to be generated just-in-time to
40 framebuffer layout in the pipeline state changes.
42 The current restrictions require there to be an exact match between the bound pipeline
49 …gCreateInfo::pColorAttachmentFormats used to create the currently bound pipeline equal to VK_FORMA…
52 …Info::depthAttachmentFormat used to create the currently bound graphics pipeline must be equal to …
[all …]
|
| D | VK_EXT_extended_dynamic_state3.adoc | 15 more pipeline state dynamic, there is still some remaining pipeline state that cannot be changed dy…
17 It was also noted that several extensions that add pipeline state lack dynamic state equivalents.
19 Reducing the amount of state required to build a pipeline can also help applications build pipeline…
21 dynamic there is less state that applications need to hash and cache for on-demand pipeline creatio…
23 These additional dynamic states can also help pare down the pipeline state to help other extensions…
24 and potentially remove the need to create and bind linked monolithic pipeline objects at all.
35 capture as much pipeline state as possible and lay a foundation for potentially doing away with lin…
36 monolithic pipeline objects.
38 Adding more and more dynamic state by itself is not the solution to all pipeline problems, and this…
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/chapters/VK_NV_device_generated_commands/ |
| D | generation.adoc | 50 pname:pipeline that this buffer memory is intended to be used with
52 * pname:pipeline is the slink:VkPipeline that this buffer memory is
67 ename:VK_PIPELINE_BIND_POINT_GRAPHICS, then pname:pipeline must: be a
75 pname:pipeline must: be a valid slink:VkPipeline handle
81 pname:pipeline must: be `NULL`
90 To bind a compute pipeline in <<device-generated-commands,Device-Generated
91 Commands>>, an application must: query the pipeline's device address.
93 [open,refpage='vkGetPipelineIndirectDeviceAddressNV',desc='Get pipeline\'s 64-bit device address',t…
97 To query a compute pipeline's 64-bit device address, call:
100 * pname:device is the logical device on which the pipeline was created.
[all …]
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/chapters/commonvalidity/ |
| D | draw_mesh_limits_common.adoc | 9 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
14 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
19 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
24 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
30 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
35 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
40 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
45 If the current pipeline bound to ename:VK_PIPELINE_BIND_POINT_GRAPHICS
|
| D | draw_vertex_binding.adoc | 29 the bound graphics pipeline state was created with the
36 If the bound graphics pipeline state was created with the
40 class>> as the pipeline
45 If the bound graphics pipeline state was created with the
51 class>> as the pipeline
56 If the bound graphics pipeline was created with both the
62 If the bound graphics pipeline was created with the
72 If the bound graphics pipeline was created with the
89 the bound graphics pipeline state was created with the
102 the bound graphics pipeline state was created with the
[all …]
|
| D | draw_dispatch_nonindirect_common.adoc | 12 pipeline bind point used by this command must: not be an unprotected
17 pipeline stages other than the framebuffer-space and compute stages in
18 the sname:VkPipeline object bound to the pipeline bind point used by
23 pipeline bind point used by this command uses the
|
| D | draw_common.adoc | 20 descriptor must: be bound to the pipeline via a descriptor set
22 If any shader executed by this pipeline accesses an code:OpTypeImage
51 set on the currently bound pipeline
57 *** there is no currently bound graphics pipeline or
59 *** the currently bound graphics pipeline was created with
78 is set on the currently bound pipeline
84 *** there is no currently bound graphics pipeline or
86 *** the currently bound graphics pipeline was created with
105 is set on the currently bound pipeline
111 *** there is no currently bound graphics pipeline or
[all …]
|
| D | draw_dispatch_common.adoc | 135 pipeline bind point, with a slink:VkPipelineLayout that is compatible
141 pipeline bind point, with a slink:VkPipelineLayout that is compatible
150 pipeline bind point, with a slink:VkPipelineLayout that is compatible
158 pipeline bind point, with a slink:VkPipelineLayout that is compatible
172 been set for the same pipeline bind point, with a slink:VkPipelineLayout
182 been set for the same pipeline bind point, with a slink:VkPipelineLayout
200 used by the slink:VkPipeline bound to the pipeline bind point used by
204 If the descriptors used by the slink:VkPipeline bound to the pipeline
211 dynamically used by the slink:VkPipeline bound to the pipeline bind
219 to the pipeline bind point used by this command
[all …]
|
| /hardware/google/gfxstream/include/OpenGLESDispatch/ |
| D | gles31_only_functions.h | 16 X(void, glBindProgramPipeline, (GLuint pipeline), (pipeline)) \
17 …X(void, glGetProgramPipelineiv, (GLuint pipeline, GLenum pname, GLint * params), (pipeline, pname,…
18 …d, glGetProgramPipelineInfoLog, (GLuint pipeline, GLsizei bufSize, GLsizei * length, GLchar * info…
19 X(void, glValidateProgramPipeline, (GLuint pipeline), (pipeline)) \
20 X(GLboolean, glIsProgramPipeline, (GLuint pipeline), (pipeline)) \
21 …X(void, glUseProgramStages, (GLuint pipeline, GLbitfield stages, GLuint program), (pipeline, stage…
22 X(void, glActiveShaderProgram, (GLuint pipeline, GLuint program), (pipeline, program)) \
|
| /hardware/google/gfxstream/host/gl/OpenGLESDispatch/ |
| D | gles31_only_dispatch_logging_wrappers.impl | 29 void glBindProgramPipeline_dispatchLoggingWrapper(GLuint pipeline) {
30 DISPATCH_DEBUG_LOG("glBindProgramPipeline(pipeline:%d)", pipeline);
31 GLDispatch::glBindProgramPipeline_underlying(pipeline);
34 void glGetProgramPipelineiv_dispatchLoggingWrapper(GLuint pipeline, GLenum pname, GLint * params) {
35 …DISPATCH_DEBUG_LOG("glGetProgramPipelineiv(pipeline:%d, pname:0x%X, params:%p)", pipeline, pname, …
36 GLDispatch::glGetProgramPipelineiv_underlying(pipeline, pname, params);
39 void glGetProgramPipelineInfoLog_dispatchLoggingWrapper(GLuint pipeline, GLsizei bufSize, GLsizei *…
40 …ISPATCH_DEBUG_LOG("glGetProgramPipelineInfoLog(pipeline:%d, bufSize:%d, length:%p, infoLog:%p)", p…
41 GLDispatch::glGetProgramPipelineInfoLog_underlying(pipeline, bufSize, length, infoLog);
44 void glValidateProgramPipeline_dispatchLoggingWrapper(GLuint pipeline) {
[all …]
|
| /hardware/google/gfxstream/codegen/vulkan/vulkan-docs-next/chapters/VK_AMD_buffer_marker/ |
| D | copies.adoc | 19 * pname:stage specifies the pipeline stage whose completion triggers the
28 pipeline stage.
31 pipeline stages occur either at the same time or earlier than this command's
39 ename:VK_ACCESS_TRANSFER_WRITE_BIT, and the pipeline stages for identifying
47 write a marker at any specific pipeline stage, it may: instead do so at any
68 pname:stage must: include only a single pipeline stage
101 the pipeline stage whose completion triggers the marker write.
109 pipeline stage.
111 fname:vkCmdWriteBufferMarkerAMD commands so long as their specified pipeline
120 ename:VK_ACCESS_TRANSFER_WRITE_BIT, and the pipeline stages for identifying
[all …]
|