Change log for August 30, 2024 Vulkan 1.3.295 spec update:

Public Issues

  * Add explicit <require> tag to XML for
    VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR in VK_KHR_pipeline_binary (public
    pull request 2423).
  * Minor wording change to avoid ambiguity of
    <<synchronization-semaphores-waiting, "`the act of waiting`">> (public
    pull request 2418).

Internal Issues

  * Deprecate VK_COLOR_SPACE_DOLBYVISION_EXT from
    VK_EXT_swapchain_colorspace and add a NOTE to VkColorSpaceKHR explaining
    why (internal issue 3884).
  * Fix markup in registry schema document to move video tag descriptions to
    the proper place and avoid mis-grouping of other tag descriptions
    (internal issue 3978).
  * Remove structextends="VkPipelineCreateInfoKHR" XML attributes from
    VkCreate*PipelineCreateInfo structs that were erroneously added with
    VK_KHR_pipeline_binaries (internal issue 3979).
  * Add vkCreateRayTracingPipelinesKHR VU and update VkPipelineCreateInfoKHR
    VU 09604 to disallow pipeline binaries with VK_NV_ray_tracing (internal
    merge request 6849).

New Extensions

  * VK_KHR_compute_shader_derivatives

ChangeLog.adoc

@@ -14,6 +14,38 @@ appears frequently in the change log.
 
 '''
 
+Change log for August 30, 2024 Vulkan 1.3.295 spec update:
+
+Public Issues
+
+  * Add explicit <require> tag to XML for
+    VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR in VK_KHR_pipeline_binary (public
+    pull request 2423).
+  * Minor wording change to avoid ambiguity of
+    <<synchronization-semaphores-waiting, "`the act of waiting`">> (public
+    pull request 2418).
+
+Internal Issues
+
+  * Deprecate VK_COLOR_SPACE_DOLBYVISION_EXT from
+    VK_EXT_swapchain_colorspace and add a NOTE to VkColorSpaceKHR explaining
+    why (internal issue 3884).
+  * Fix markup in registry schema document to move video tag descriptions to
+    the proper place and avoid mis-grouping of other tag descriptions
+    (internal issue 3978).
+  * Remove structextends="VkPipelineCreateInfoKHR" XML attributes from
+    VkCreate*PipelineCreateInfo structs that were erroneously added with
+    VK_KHR_pipeline_binaries (internal issue 3979).
+  * Add vkCreateRayTracingPipelinesKHR VU and update VkPipelineCreateInfoKHR
+    VU 09604 to disallow pipeline binaries with VK_NV_ray_tracing (internal
+    merge request 6849).
+
+New Extensions
+
+  * VK_KHR_compute_shader_derivatives
+
+'''
+
 Change log for August 23, 2024 Vulkan 1.3.294 spec update:
 
 Public Issues

Makefile

@@ -139,7 +139,7 @@ VERBOSE =
 # ADOCOPTS options for asciidoc->HTML5 output
 
 NOTEOPTS     = -a editing-notes -a implementation-guide
-PATCHVERSION = 294
+PATCHVERSION = 295
 BASEOPTS     =
 
 ifneq (,$(findstring VKSC_VERSION_1_0,$(VERSIONS)))

appendices/VK_KHR_compute_shader_derivatives.adoc

@@ -0,0 +1,71 @@
+// Copyright 2020-2024 The Khronos Group Inc.
+//
+// SPDX-License-Identifier: CC-BY-4.0
+
+include::{generated}/meta/{refprefix}VK_KHR_compute_shader_derivatives.adoc[]
+
+=== Other Extension Metadata
+
+*Last Modified Date*::
+    2024-06-26
+*IP Status*::
+    No known IP claims.
+*Interactions and External Dependencies*::
+  - This extension requires
+    {spirv}/KHR/SPV_KHR_compute_shader_derivatives.html[`SPV_KHR_compute_shader_derivatives`]
+  - This extension provides API support for
+    {GLSLregistry}/khr/GLSL_KHR_compute_shader_derivatives.txt[`GL_KHR_compute_shader_derivatives`]
+*Contributors*::
+  - Jean-Noe Morissette, Epic Games
+  - Daniel Koch, NVIDIA
+  - Pat Brown, NVIDIA
+  - Stu Smith, AMD
+  - Jan-Harald Fredriksen, Arm
+  - Tobias Hector, AMD
+  - Ralph Potter, Samsung
+  - Pan Gao, Huawei
+  - Samuel (Sheng-Wen) Huang, MediaTek
+  - Graeme Leese, Broadcom
+  - Hans-Kristian Arntzen, Valve
+  - Matthew Netsh, Qualcomm
+
+=== Description
+
+This extension adds Vulkan support for the
+{spirv}/KHR/SPV_KHR_compute_shader_derivatives.html[`SPV_KHR_compute_shader_derivatives`]
+SPIR-V extension.
+
+The SPIR-V extension provides two new execution modes, both of which allow
+execution models with defined workgroups to use built-ins that evaluate
+derivatives explicitly or implicitly.
+Derivatives will be computed via differencing over a 2x2 group of shader
+invocations.
+The code:DerivativeGroupQuadsKHR execution mode assembles shader invocations
+into 2x2 groups, where each group has x and y coordinates of the local
+invocation ID of the form (2m+{0,1}, 2n+{0,1}).
+The code:DerivativeGroupLinearKHR execution mode assembles shader
+invocations into 2x2 groups, where each group has local invocation index
+values of the form 4m+{0,1,2,3}.
+
+The new execution modes are supported in compute shaders and optionally (see
+<<limits-meshAndTaskShaderDerivatives, meshAndTaskShaderDerivatives>>) in
+mesh and task shaders.
+
+include::{generated}/interfaces/VK_KHR_compute_shader_derivatives.adoc[]
+
+=== New SPIR-V Capability
+
+  * <<spirvenv-capabilities-table-ComputeDerivativeGroupQuadsKHR,
+    code:ComputeDerivativeGroupQuadsKHR>>
+  * <<spirvenv-capabilities-table-ComputeDerivativeGroupLinearKHR,
+    code:ComputeDerivativeGroupLinearKHR>>
+
+=== Examples
+
+None.
+
+=== Version History
+
+  * Revision 1, 2023-02-27 (Jean-Noe Morissette)
+  ** Initial draft
+  ** Add properties and clarify mesh and task support (Daniel Koch)

appendices/glossary.adoc

@@ -412,12 +412,12 @@ Depth/Stencil Resolve Attachment::
 endif::VK_VERSION_1_2,VK_KHR_depth_stencil_resolve[]
 
 Derivative Group::
-    A set of fragment
-ifdef::VK_NV_compute_shader_derivatives[]
-    or compute
-endif::VK_NV_compute_shader_derivatives[]
-    shader invocations that cooperate to compute derivatives, including
-    implicit derivatives for sampled image operations.
+    A set of
+ifdef::VK_KHR_compute_shader_derivatives,VK_NV_compute_shader_derivatives[compute,]
+ifdef::VK_KHR_compute_shader_derivatives[mesh, task,]
+ifdef::VK_KHR_compute_shader_derivatives,VK_NV_compute_shader_derivatives[or]
+    fragment shader invocations that cooperate to compute derivatives,
+    including implicit derivatives for sampled image operations.
 
 Descriptor::
     Information about a resource or resource view written into a descriptor

appendices/spirvenv.adoc

@@ -489,6 +489,14 @@ or knowledge of runtime information, such as enabled features.
     For compute shaders using the code:DerivativeGroupLinearNV execution
     mode, the product of the dimensions of the local workgroup size must: be
     a multiple of four
+  * [[VUID-{refpage}-DerivativeGroupQuadsKHR-10151]]
+    For compute, mesh, or task shaders using the
+    code:DerivativeGroupQuadsKHR execution mode, the first two dimensions of
+    the local workgroup size must: be a multiple of two
+  * [[VUID-{refpage}-DerivativeGroupLinearKHR-10152]]
+    For compute, mesh, or task shaders using the
+    code:DerivativeGroupLinearKHR execution mode, the product of the
+    dimensions of the local workgroup size must: be a multiple of four
   * [[VUID-{refpage}-OpGroupNonUniformBallotBitCount-04685]]
     If code:OpGroupNonUniformBallotBitCount is used, the group operation
     must: be limited to code:Reduce, code:InclusiveScan, or
@@ -980,7 +988,6 @@ ifdef::VK_VERSION_1_1[]
     buffer, the storage texel buffer through which that memory is accessed
     must: be dynamically uniform within the invocation group
 endif::VK_VERSION_1_1[]
-  * [[VUID-{refpage}-NonUniform-06274]]
   * [[VUID-{refpage}-None-10148]]
     If an instruction accesses memory through any resource,
 ifdef::VK_VERSION_1_1[]
@@ -2255,6 +2262,14 @@ endif::VK_KHR_shader_subgroup_rotate[]
     If <<limits-protectedNoFault, pname:protectedNoFault>> is not supported,
     the {StorageClass} of the code:PhysicalStorageBuffer must: not be used
     if the buffer being accessed is <<memory-protected-memory,protected>>
+ifdef::VK_KHR_compute_shader_derivatives[]
+  * [[VUID-{refpage}-meshAndTaskShaderDerivatives-10153]]
+    If <<limits-meshAndTaskShaderDerivatives,
+    pname:meshAndTaskShaderDerivatives>> is ename:VK_FALSE, the
+    code:DerivativeGroupLinearKHR and code:DerivativeGroupQuadsKHR execution
+    modes must: not be used in the code:MeshEXT, code:MeshNV, code:TaskEXT,
+    or code:TaskNV {ExecutionModel}
+endif::VK_KHR_compute_shader_derivatives[]
 ****
 --
 

chapters/VK_KHR_surface/wsi.adoc

@@ -1302,11 +1302,6 @@ ifdef::VK_EXT_swapchain_colorspace[]
   * ename:VK_COLOR_SPACE_HDR10_ST2084_EXT specifies support for the images
     in HDR10 (BT2020) color space, encoded according to SMPTE ST2084
     Perceptual Quantizer (PQ) specification.
-  * ename:VK_COLOR_SPACE_DOLBYVISION_EXT specifies support for the images in
-    Dolby Vision (BT2020) color space, encoded according to SMPTE ST2084
-    Perceptual Quantizer (PQ) specification.
-    The presentation engine is expected to use Dolby's proprietary
-    techniques to display the image.
   * ename:VK_COLOR_SPACE_HDR10_HLG_EXT specifies support for the images in
     HDR10 (BT2020) color space, encoded according to the Hybrid Log Gamma
     (HLG) specification.
@@ -1332,6 +1327,8 @@ ifdef::hidden[]
   * elink:VkColorSpaceKHR (deprecated aliases)
   ** etext:VK_COLORSPACE_SRGB_NONLINEAR_KHR <<SCID-8>>
   ** etext:VK_COLOR_SPACE_DCI_P3_LINEAR_EXT <<SCID-8>>
+  * elink:VkColorSpaceKHR (not well-defined)
+  ** ename:VK_COLOR_SPACE_DOLBYVISION_EXT <<SCID-8>>
 // end::scremoved[]
 endif::hidden[]
 endif::VKSC_VERSION_1_0[]
@@ -1361,6 +1358,20 @@ endif::VKSC_VERSION_1_0[]
 
 [NOTE]
 ====
+In older versions of the `apiext:VK_EXT_swapchain_colorspace` extension,
+ename:VK_COLOR_SPACE_DOLBYVISION_EXT was exposed.
+The intent was to indicate the presentation engine shall decode an image
+using the SMPTE ST 2084 Perceptual Quantizer (PQ) EOTF, and then apply a
+proprietary OOTF to process the image.
+However, Dolby Vision profile 8.4 describes an encoding using the Hybrid Log
+Gamma (HLG) OETF, and there is no swapchain extension for signaling Dolby
+Vision metadata to be used by a proprietary OOTF.
+This enum is deprecated but is maintained for backwards compatibility.
+====
+
+[NOTE]
+.Note
+====
 For a traditional "`Linear`" or non-gamma transfer function color space use
 ename:VK_COLOR_SPACE_PASS_THROUGH_EXT.
 ====
@@ -1392,7 +1403,6 @@ the following color spaces:
 | sRGB           | 0.640, 0.330 | 0.300, 0.600  | 0.150, 0.060 | 0.3127, 0.3290 (D65) | sRGB
 | extended sRGB  | 0.640, 0.330 | 0.300, 0.600  | 0.150, 0.060 | 0.3127, 0.3290 (D65) | scRGB
 | HDR10_ST2084   | 0.708, 0.292 | 0.170, 0.797  | 0.131, 0.046 | 0.3127, 0.3290 (D65) | ST2084 PQ
-| DOLBYVISION    | 0.708, 0.292 | 0.170, 0.797  | 0.131, 0.046 | 0.3127, 0.3290 (D65) | ST2084 PQ
 | HDR10_HLG      | 0.708, 0.292 | 0.170, 0.797  | 0.131, 0.046 | 0.3127, 0.3290 (D65) | HLG
 | Adobe RGB      | 0.640, 0.330 | 0.210, 0.710  | 0.150, 0.060 | 0.3127, 0.3290 (D65) | Adobe RGB
 |====

chapters/features.adoc

@@ -2525,13 +2525,19 @@ include::{generated}/validity/structs/VkPhysicalDeviceCornerSampledImageFeatures
 --
 endif::VK_NV_corner_sampled_image[]
 
-ifdef::VK_NV_compute_shader_derivatives[]
-[open,refpage='VkPhysicalDeviceComputeShaderDerivativesFeaturesNV',desc='Structure describing compute shader derivative features that can be supported by an implementation',type='structs']
+ifdef::VK_KHR_compute_shader_derivatives,VK_NV_compute_shader_derivatives[]
+[open,refpage='VkPhysicalDeviceComputeShaderDerivativesFeaturesKHR',desc='Structure describing compute shader derivative features that can be supported by an implementation',type='structs',alias="VkPhysicalDeviceComputeShaderDerivativesFeaturesNV"]
 --
-The sname:VkPhysicalDeviceComputeShaderDerivativesFeaturesNV structure is
+The sname:VkPhysicalDeviceComputeShaderDerivativesFeaturesKHR structure is
 defined as:
 
+include::{generated}/api/structs/VkPhysicalDeviceComputeShaderDerivativesFeaturesKHR.adoc[]
+
+ifdef::VK_NV_compute_shader_derivatives[]
+or the equivalent
+
 include::{generated}/api/structs/VkPhysicalDeviceComputeShaderDerivativesFeaturesNV.adoc[]
+endif::VK_NV_compute_shader_derivatives[]
 
 This structure describes the following features:
 
@@ -2540,19 +2546,19 @@ This structure describes the following features:
     structure.
   * [[features-computeDerivativeGroupQuads]]
     pname:computeDerivativeGroupQuads indicates that the implementation
-    supports the code:ComputeDerivativeGroupQuadsNV SPIR-V capability.
+    supports the code:ComputeDerivativeGroupQuadsKHR SPIR-V capability.
   * [[features-computeDerivativeGroupLinear]]
     pname:computeDerivativeGroupLinear indicates that the implementation
-    supports the code:ComputeDerivativeGroupLinearNV SPIR-V capability.
+    supports the code:ComputeDerivativeGroupLinearKHR SPIR-V capability.
 
 See <<shaders-scope-quad, Quad shader scope>> for more information.
 
-:refpage: VkPhysicalDeviceComputeShaderDerivativesFeaturesNVfeatures.
+:refpage: VkPhysicalDeviceComputeShaderDerivativesFeaturesKHR.
 include::{chapters}/features.adoc[tag=features]
 
-include::{generated}/validity/structs/VkPhysicalDeviceComputeShaderDerivativesFeaturesNV.adoc[]
+include::{generated}/validity/structs/VkPhysicalDeviceComputeShaderDerivativesFeaturesKHR.adoc[]
 --
-endif::VK_NV_compute_shader_derivatives[]
+endif::VK_KHR_compute_shader_derivatives,VK_NV_compute_shader_derivatives[]
 
 ifdef::VK_KHR_fragment_shader_barycentric[]
 [open,refpage='VkPhysicalDeviceFragmentShaderBarycentricFeaturesKHR',desc='Structure describing barycentric support in fragment shaders that can be supported by an implementation',type='structs']
@@ -8113,6 +8119,11 @@ ifdef::VK_KHR_shader_float16_int8[]
     <<features-shaderInt8,pname:shaderInt8>> if
     `apiext:VK_KHR_shader_float16_int8` is supported.
 endif::VK_KHR_shader_float16_int8[]
+ifdef::VK_KHR_compute_shader_derivatives[]
+  * <<features-computeDerivativeGroupLinear,
+    pname:computeDerivativeGroupLinear>>, if the
+    `apiext:VK_KHR_compute_shader_derivatives` extension is supported.
+endif::VK_KHR_compute_shader_derivatives[]
 
 ifdef::VKSC_VERSION_1_0[]
 ifdef::hidden[]

chapters/limits.adoc

@@ -4866,6 +4866,30 @@ include::{generated}/validity/structs/VkPhysicalDeviceRenderPassStripedPropertie
 
 endif::VK_ARM_render_pass_striped[]
 
+ifdef::VK_KHR_compute_shader_derivatives[]
+[open,refpage='VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR',desc='Structure describing compute shader derivative operations supported by an implementation',type='structs']
+--
+The sname:VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR structure is
+defined as:
+
+include::{generated}/api/structs/VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR.adoc[]
+
+The members of the
+sname:VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR structure
+describe the following:
+
+  * [[limits-meshAndTaskShaderDerivatives]]
+    pname:meshAndTaskShaderDerivatives indicates whether the mesh and task
+    shader stages support the code:ComputeDerivativeGroupQuadsKHR and
+    code:ComputeDerivativeGroupLinearKHR SPIR-V capabilities.
+
+:refpage: VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR
+include::{chapters}/limits.adoc[tag=limits_desc]
+
+include::{generated}/validity/structs/VkPhysicalDeviceComputeShaderDerivativesPropertiesKHR.adoc[]
+--
+endif::VK_KHR_compute_shader_derivatives[]
+
 ifdef::VK_EXT_map_memory_placed[]
 [open,refpage='VkPhysicalDeviceMapMemoryPlacedPropertiesEXT',desc='Structure describing the alignment requirements of placed memory maps for a physical device',type='structs']
 --

chapters/pipelines.adoc

@@ -6043,6 +6043,10 @@ endif::VKSC_VERSION_1_0[]
 ****
 include::{chapters}/commonvalidity/create_ray_tracing_pipelines_common.adoc[]
 include::{chapters}/commonvalidity/pipeline_create_common.adoc[]
+  * [[VUID-vkCreateRayTracingPipelinesNV-pNext-10150]]
+    If a slink:VkPipelineCreateFlags2CreateInfoKHR structure is included in
+    the pname:pNext chain of any element of pname:pCreateInfos,
+    ename:VK_PIPELINE_CREATE_2_CAPTURE_DATA_BIT_KHR flag must: not be set
 ****
 ifdef::VKSC_VERSION_1_0[]
 ifdef::hidden[]
@@ -8188,7 +8192,6 @@ include::{generated}/api/structs/VkPipelineCreateInfoKHR.adoc[]
     pname:pNext must: be pointer to a valid instance of
     slink:VkGraphicsPipelineCreateInfo,
 ifdef::VK_AMDX_shader_enqueue[slink:VkExecutionGraphPipelineCreateInfoAMDX,]
-ifdef::VK_NV_ray_tracing[slink:VkRayTracingPipelineCreateInfoNV,]
 ifdef::VK_KHR_ray_tracing_pipeline[slink:VkRayTracingPipelineCreateInfoKHR,]
     or slink:VkComputePipelineCreateInfo
 ****
@@ -8226,9 +8229,8 @@ include::{generated}/validity/structs/VkPipelineBinaryKeyKHR.adoc[]
 
 [open,refpage='VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR',desc='Maximum length of binary key',type='consts']
 --
-ename:VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR is the length in bytes of a
-binary key, as returned in
-slink:VkPipelineBinaryKeyKHR::pname:keySize.
+ename:VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR is the length in bytes of a binary
+key, as returned in slink:VkPipelineBinaryKeyKHR::pname:keySize.
 
 include::{generated}/api/enums/VK_MAX_PIPELINE_BINARY_KEY_SIZE_KHR.adoc[]
 --