// Copyright (c) 2015-2019 Khronos Group. This work is licensed under a // Creative Commons Attribution 4.0 International License; see // http://creativecommons.org/licenses/by/4.0/ [appendix] [[spirvenv]] = Vulkan Environment for SPIR-V Shaders for Vulkan are defined by the <> as well as the <> Specification. This appendix defines additional SPIR-V requirements applying to Vulkan shaders. == Versions and Formats ifdef::VK_VERSION_1_1[] A Vulkan 1.1 implementation must: support the 1.0, 1.1, 1.2, and 1.3 versions of SPIR-V and the 1.0 version of the SPIR-V Extended Instructions for GLSL. endif::VK_VERSION_1_1[] ifndef::VK_VERSION_1_1[] A Vulkan 1.0 implementation must: support the 1.0 version of SPIR-V and the 1.0 version of the SPIR-V Extended Instructions for GLSL. endif::VK_VERSION_1_1[] A SPIR-V module passed into flink:vkCreateShaderModule is interpreted as a series of 32-bit words in host endianness, with literal strings packed as described in section 2.2 of the SPIR-V Specification. The first few words of the SPIR-V module must: be a magic number and a SPIR-V version number, as described in section 2.3 of the SPIR-V Specification. [[spirvenv-capabilities]] == Capabilities The SPIR-V capabilities listed below must: be supported if the corresponding feature or extension is enabled, or if no features or extensions are listed for that capability. Extensions are only listed when there is not also a feature bit associated with that capability. [[spirvenv-capabilities-table]] .List of SPIR-V Capabilities and enabling features or extensions [options="header"] |==== | SPIR-V code:OpCapability | Vulkan feature or extension name | code:Matrix | | code:Shader | | code:InputAttachment | | code:Sampled1D | | code:Image1D | | code:SampledBuffer | | code:ImageBuffer | | code:ImageQuery | | code:DerivativeControl | | code:Geometry | <> | code:Tessellation | <> | code:Float64 | <> | code:Int64 | <> ifdef::VK_KHR_shader_atomic_int64[] [[spirvenv-capabilities-table-int64atomics]] | code:Int64Atomics | <> endif::VK_KHR_shader_atomic_int64[] | code:Int16 | <> | code:TessellationPointSize | <> | code:GeometryPointSize | <> | code:ImageGatherExtended | <> | code:StorageImageMultisample | <> | code:UniformBufferArrayDynamicIndexing | <> | code:SampledImageArrayDynamicIndexing | <> | code:StorageBufferArrayDynamicIndexing | <> | code:StorageImageArrayDynamicIndexing | <> | code:ClipDistance | <> | code:CullDistance | <> | code:ImageCubeArray | <> | code:SampleRateShading | <> | code:SparseResidency | <> | code:MinLod | <> | code:SampledCubeArray | <> | code:ImageMSArray | <> | code:StorageImageExtendedFormats | | code:InterpolationFunction | <> | code:StorageImageReadWithoutFormat | <> | code:StorageImageWriteWithoutFormat | <> | code:MultiViewport | <> ifdef::VK_VERSION_1_1,VK_KHR_shader_draw_parameters[] | code:DrawParameters | ifdef::VK_VERSION_1_1[] <> endif::VK_VERSION_1_1[] ifdef::VK_KHR_shader_draw_parameters+VK_VERSION_1_1[] or endif::VK_KHR_shader_draw_parameters+VK_VERSION_1_1[] ifdef::VK_KHR_shader_draw_parameters[] <> endif::VK_KHR_shader_draw_parameters[] endif::VK_VERSION_1_1,VK_KHR_shader_draw_parameters[] ifdef::VK_VERSION_1_1,VK_KHR_multiview[] [[spirvenv-capabilities-multiview]] | code:MultiView | ifndef::VK_VERSION_1_1[] <> endif::VK_VERSION_1_1[] endif::VK_VERSION_1_1,VK_KHR_multiview[] ifdef::VK_VERSION_1_1,VK_KHR_device_group[] | code:DeviceGroup | ifndef::VK_VERSION_1_1[] <> endif::VK_VERSION_1_1[] endif::VK_VERSION_1_1,VK_KHR_device_group[] ifdef::VK_VERSION_1_1,VK_KHR_variable_pointers[] [[spirvenv-capabilities-table-variablepointers]] | code:VariablePointersStorageBuffer | <> | code:VariablePointers | <> endif::VK_VERSION_1_1,VK_KHR_variable_pointers[] ifdef::VK_EXT_shader_stencil_export[] [[spirvenv-capabilities-table-shaderstencilexportext]] | code:StencilExportEXT | `<>` endif::VK_EXT_shader_stencil_export[] ifdef::VK_EXT_shader_subgroup_ballot[] [[spirvenv-capabilities-table-subgroupballot]] | code:SubgroupBallotKHR | `<>` endif::VK_EXT_shader_subgroup_ballot[] ifdef::VK_EXT_shader_subgroup_vote[] [[spirvenv-capabilities-table-subgroupvote]] | code:SubgroupVoteKHR | `<>` endif::VK_EXT_shader_subgroup_vote[] ifdef::VK_AMD_shader_image_load_store_lod[] [[spirvenv-capabilities-table-imagereadwritelodamd]] | code:ImageReadWriteLodAMD | `<>` endif::VK_AMD_shader_image_load_store_lod[] ifdef::VK_AMD_texture_gather_bias_lod[] [[spirvenv-capabilities-table-imagegatherbiaslodamd]] | code:ImageGatherBiasLodAMD | `<>` endif::VK_AMD_texture_gather_bias_lod[] ifdef::VK_AMD_shader_fragment_mask[] [[spirvenv-capabilities-table-fragmentmaskamd]] | code:FragmentMaskAMD | `<>` endif::VK_AMD_shader_fragment_mask[] ifdef::VK_NV_sample_mask_override_coverage[] [[spirvenv-capabilities-table-samplemaskoverridecoverage]] | code:SampleMaskOverrideCoverageNV | `<>` endif::VK_NV_sample_mask_override_coverage[] ifdef::VK_NV_geometry_shader_passthrough[] [[spirvenv-capabilities-table-geometryshaderpassthrough]] | code:GeometryShaderPassthroughNV | `<>` endif::VK_NV_geometry_shader_passthrough[] ifdef::VK_EXT_shader_viewport_index_layer[] [[spirvenv-capabilities-table-shader-viewport-index-layer]] | code:ShaderViewportIndexLayerEXT | `<>` endif::VK_EXT_shader_viewport_index_layer[] ifdef::VK_NV_viewport_array2[] [[spirvenv-capabilities-table-viewportarray2]] | code:ShaderViewportIndexLayerNV | `<>` | code:ShaderViewportMaskNV | `<>` endif::VK_NV_viewport_array2[] ifdef::VK_NVX_multiview_per_view_attributes[] [[spirvenv-capabilities-table-perviewattributes]] | code:PerViewAttributesNV | `<>` endif::VK_NVX_multiview_per_view_attributes[] ifdef::VK_VERSION_1_1,VK_KHR_16bit_storage[] [[spirvenv-capabilities-table-16bitstorage]] | code:StorageBuffer16BitAccess | <> | code:UniformAndStorageBuffer16BitAccess | <> | code:StoragePushConstant16 | <> | code:StorageInputOutput16 | <> endif::VK_VERSION_1_1,VK_KHR_16bit_storage[] ifdef::VK_VERSION_1_1[] [[spirvenv-capabilities-table-subgroup]] | code:GroupNonUniform | <> | code:GroupNonUniformVote | <> | code:GroupNonUniformArithmetic | <> | code:GroupNonUniformBallot | <> | code:GroupNonUniformShuffle | <> | code:GroupNonUniformShuffleRelative | <> | code:GroupNonUniformClustered | <> | code:GroupNonUniformQuad | <> ifdef::VK_NV_shader_subgroup_partitioned[] | code:GroupNonUniformPartitionedNV | <> endif::VK_NV_shader_subgroup_partitioned[] endif::VK_VERSION_1_1[] ifdef::VK_EXT_post_depth_coverage[] [[spirvenv-capabilities-table-postdepthcoverage]] | code:SampleMaskPostDepthCoverage | `<>` endif::VK_EXT_post_depth_coverage[] ifdef::VK_EXT_descriptor_indexing[] [[spirvenv-capabilities-table-descriptorindexing]] | code:ShaderNonUniformEXT | `<>` | code:RuntimeDescriptorArrayEXT | <> | code:InputAttachmentArrayDynamicIndexingEXT | <> | code:UniformTexelBufferArrayDynamicIndexingEXT | <> | code:StorageTexelBufferArrayDynamicIndexingEXT | <> | code:UniformBufferArrayNonUniformIndexingEXT | <> | code:SampledImageArrayNonUniformIndexingEXT | <> | code:StorageBufferArrayNonUniformIndexingEXT | <> | code:StorageImageArrayNonUniformIndexingEXT | <> | code:InputAttachmentArrayNonUniformIndexingEXT | <> | code:UniformTexelBufferArrayNonUniformIndexingEXT | <> | code:StorageTexelBufferArrayNonUniformIndexingEXT | <> endif::VK_EXT_descriptor_indexing[] ifdef::VK_KHR_shader_float16_int8,VK_AMD_gpu_shader_half_float[] | code:Float16 | ifdef::VK_KHR_shader_float16_int8[] <> endif::VK_KHR_shader_float16_int8[] ifdef::VK_KHR_shader_float16_int8+VK_AMD_gpu_shader_half_float[or] ifdef::VK_AMD_gpu_shader_half_float[] `<>` endif::VK_AMD_gpu_shader_half_float[] endif::VK_KHR_shader_float16_int8,VK_AMD_gpu_shader_half_float[] ifdef::VK_KHR_shader_float16_int8[] | code:Int8 | <> endif::VK_KHR_shader_float16_int8[] ifdef::VK_KHR_8bit_storage[] [[spirvenv-capabilities-table-8bitstorage]] | code:StorageBuffer8BitAccess | <> | code:UniformAndStorageBuffer8BitAccess | <> | code:StoragePushConstant8 | <> endif::VK_KHR_8bit_storage[] ifdef::VK_KHR_vulkan_memory_model[] [[spirvenv-capabilities-table-memorymodel]] | code:VulkanMemoryModelKHR | <> | code:VulkanMemoryModelDeviceScopeKHR | <> endif::VK_KHR_vulkan_memory_model[] ifdef::VK_KHR_shader_float_controls[] [[spirvenv-capabilities-table-shaderfloatcontrols]] | code:DenormPreserve | <>, <>, <> | code:DenormFlushToZero | <>, <>, <> | code:SignedZeroInfNanPreserve | <>, <>, <> | code:RoundingModeRTE | <>, <>, <> | code:RoundingModeRTZ | <>, <>, <> endif::VK_KHR_shader_float_controls[] ifdef::VK_NV_compute_shader_derivatives[] [[spirvenv-capabilities-table-computederivatives-quads]] | code:ComputeDerivativeGroupQuadsNV | <> [[spirvenv-capabilities-table-computederivatives-linear]] | code:ComputeDerivativeGroupLinearNV | <> endif::VK_NV_compute_shader_derivatives[] ifdef::VK_NV_fragment_shader_barycentric[] [[spirvenv-capabilities-fragment-barycentric]] | code:FragmentBarycentricNV | <> endif::VK_NV_fragment_shader_barycentric[] ifdef::VK_NV_shader_image_footprint[] [[spirvenv-capabilities-table-imagefootprint]] | code:ImageFootprintNV | <> endif::VK_NV_shader_image_footprint[] ifdef::VK_NV_shading_rate_image[] | code:ShadingRateImageNV | <> endif::VK_NV_shading_rate_image[] ifdef::VK_NV_mesh_shader[] [[spirvenv-capabilities-table-meshshading]] | code:MeshShadingNV | `<>` endif::VK_NV_mesh_shader[] ifdef::VK_NV_ray_tracing[] [[spirvenv-capabilities-table-raytracing]] | code:RayTracingNV | `<>` endif::VK_NV_ray_tracing[] ifdef::VK_EXT_transform_feedback[] | code:TransformFeedback | <> | code:GeometryStreams | <> endif::VK_EXT_transform_feedback[] ifdef::VK_EXT_fragment_density_map[] [[spirvenv-capabilities-table-fragmentdensity]] | code:FragmentDensityEXT | <> endif::VK_EXT_fragment_density_map[] ifdef::VK_EXT_buffer_device_address[] [[spirvenv-capabilities-table-physicalstoragebufferaddresses]] | code:PhysicalStorageBufferAddressesEXT | <> endif::VK_EXT_buffer_device_address[] ifdef::VK_NV_cooperative_matrix[] [[spirvenv-capabilities-table-cooperativeMatrix]] | code:CooperativeMatrixNV | <> endif::VK_NV_cooperative_matrix[] |==== ifdef::VK_VERSION_1_1,VK_KHR_variable_pointers[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_variable_pointers` SPIR-V extension. endif::VK_VERSION_1_1,VK_KHR_variable_pointers[] ifdef::VK_AMD_shader_explicit_vertex_parameter[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_shader_explicit_vertex_parameter` SPIR-V extension. endif::VK_AMD_shader_explicit_vertex_parameter[] ifdef::VK_AMD_gcn_shader[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_gcn_shader` SPIR-V extension. endif::VK_AMD_gcn_shader[] ifdef::VK_AMD_gpu_shader_half_float[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_gpu_shader_half_float` SPIR-V extension. endif::VK_AMD_gpu_shader_half_float[] ifdef::VK_AMD_gpu_shader_int16[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_gpu_shader_int16` SPIR-V extension. endif::VK_AMD_gpu_shader_int16[] ifdef::VK_AMD_shader_ballot[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_shader_ballot` SPIR-V extension. endif::VK_AMD_shader_ballot[] ifdef::VK_AMD_shader_fragment_mask[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_shader_fragment_mask` SPIR-V extension. endif::VK_AMD_shader_fragment_mask[] ifdef::VK_AMD_shader_image_load_store_lod[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_shader_image_load_store_lod` SPIR-V extension. endif::VK_AMD_shader_image_load_store_lod[] ifdef::VK_AMD_shader_trinary_minmax[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_shader_trinary_minmax` SPIR-V extension. endif::VK_AMD_shader_trinary_minmax[] ifdef::VK_AMD_texture_gather_bias_lod[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_AMD_texture_gather_bias_lod` SPIR-V extension. endif::VK_AMD_texture_gather_bias_lod[] ifdef::VK_VERSION_1_1,VK_KHR_shader_draw_parameters[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_shader_draw_parameters` SPIR-V extension. endif::VK_VERSION_1_1,VK_KHR_shader_draw_parameters[] ifdef::VK_KHR_8bit_storage[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_8bit_storage` SPIR-V extension. endif::VK_KHR_8bit_storage[] ifdef::VK_VERSION_1_1,VK_KHR_16bit_storage[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the https://www.khronos.org/registry/spir-v/extensions/KHR/SPV_KHR_16bit_storage.html[`SPV_KHR_16bit_storage`] SPIR-V extension. endif::VK_VERSION_1_1,VK_KHR_16bit_storage[] ifdef::VK_KHR_shader_float_controls[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the https://www.khronos.org/registry/spir-v/extensions/KHR/SPV_KHR_float_controls.html[`SPV_KHR_float_controls`] SPIR-V extension. endif::VK_KHR_shader_float_controls[] ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the https://www.khronos.org/registry/spir-v/extensions/KHR/SPV_KHR_storage_buffer_storage_class.html[`SPV_KHR_storage_buffer_storage_class`] SPIR-V extension. endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ifdef::VK_EXT_post_depth_coverage[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_post_depth_coverage` SPIR-V extension. endif::VK_EXT_post_depth_coverage[] ifdef::VK_EXT_shader_stencil_export[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_EXT_shader_stencil_export` SPIR-V extension. endif::VK_EXT_shader_stencil_export[] ifdef::VK_EXT_shader_subgroup_ballot[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_shader_ballot` SPIR-V extension. endif::VK_EXT_shader_subgroup_ballot[] ifdef::VK_EXT_shader_subgroup_vote[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_subgroup_vote` SPIR-V extension. endif::VK_EXT_shader_subgroup_vote[] ifdef::VK_NV_sample_mask_override_coverage[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_sample_mask_override_coverage` SPIR-V extension. endif::VK_NV_sample_mask_override_coverage[] ifdef::VK_NV_geometry_shader_passthrough[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_geometry_shader_passthrough` SPIR-V extension. endif::VK_NV_geometry_shader_passthrough[] ifdef::VK_NV_mesh_shader[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_mesh_shader` SPIR-V extension. endif::VK_NV_mesh_shader[] ifdef::VK_NV_viewport_array2[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_viewport_array2` SPIR-V extension. endif::VK_NV_viewport_array2[] ifdef::VK_EXT_shader_viewport_index_layer[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_EXT_shader_viewport_index_layer` SPIR-V extension. endif::VK_EXT_shader_viewport_index_layer[] ifdef::VK_NVX_multiview_per_view_attributes[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NVX_multiview_per_view_attributes` SPIR-V extension. endif::VK_NVX_multiview_per_view_attributes[] ifdef::VK_EXT_descriptor_indexing[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_EXT_descriptor_indexing` SPIR-V extension. endif::VK_EXT_descriptor_indexing[] ifdef::VK_KHR_vulkan_memory_model[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_KHR_vulkan_memory_model` SPIR-V extension. endif::VK_KHR_vulkan_memory_model[] ifdef::VK_NV_compute_shader_derivatives[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_compute_shader_derivatives` SPIR-V extension. endif::VK_NV_compute_shader_derivatives[] ifdef::VK_NV_fragment_shader_barycentric[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_fragment_shader_barycentric` SPIR-V extension. endif::VK_NV_fragment_shader_barycentric[] ifdef::VK_NV_shader_image_footprint[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_shader_image_footprint` SPIR-V extension. endif::VK_NV_shader_image_footprint[] ifdef::VK_NV_shading_rate_image[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_shading_rate` SPIR-V extension. endif::VK_NV_shading_rate_image[] ifdef::VK_NV_ray_tracing[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_ray_tracing` SPIR-V extension. endif::VK_NV_ray_tracing[] ifdef::VK_GOOGLE_hlsl_functionality1[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_GOOGLE_hlsl_functionality1` SPIR-V extension. endif::VK_GOOGLE_hlsl_functionality1[] ifdef::VK_GOOGLE_decorate_string[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_GOOGLE_decorate_string` SPIR-V extension. endif::VK_GOOGLE_decorate_string[] ifdef::VK_EXT_fragment_density_map[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_EXT_fragment_invocation_density` SPIR-V extension. endif::VK_EXT_fragment_density_map[] ifdef::VK_EXT_buffer_device_address[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_EXT_physical_storage_buffer` SPIR-V extension. endif::VK_EXT_buffer_device_address[] ifdef::VK_NV_cooperative_matrix[] The application can: pass a SPIR-V module to flink:vkCreateShaderModule that uses the `SPV_NV_cooperative_matrix` SPIR-V extension. endif::VK_NV_cooperative_matrix[] The application must: not pass a SPIR-V module containing any of the following to flink:vkCreateShaderModule: * any OpCapability not listed above, * an unsupported capability, or * a capability which corresponds to a Vulkan feature or extension which has not been enabled. [[spirvenv-module-validation]] == Validation Rules within a Module A SPIR-V module passed to flink:vkCreateShaderModule must: conform to the following rules: * Every entry point must: have no return value and accept no arguments. * Recursion: The static function-call graph for an entry point must: not contain cycles. * The *Logical* addressing model must: be selected. * *Scope* for execution must: be limited to: ** *Workgroup* ** *Subgroup* * *Scope* for memory must: be limited to: ** *Device* ifdef::VK_KHR_vulkan_memory_model[] *** If <> is enabled and <> is not enabled, *Device* scope must: not be used. *** If <> is not enabled, *Device* scope only extends to the queue family, not the whole device. endif::VK_KHR_vulkan_memory_model[] ifndef::VK_KHR_vulkan_memory_model[] *** *Device* scope only extends to the queue family, not the whole device. endif::VK_KHR_vulkan_memory_model[] ifdef::VK_KHR_vulkan_memory_model[] ** *QueueFamilyKHR* *** If <> is not enabled, *QueueFamilyKHR* must: not be used. endif::VK_KHR_vulkan_memory_model[] ** *Workgroup* ifdef::VK_VERSION_1_1[] ** *Subgroup* endif::VK_VERSION_1_1[] ** *Invocation* ifdef::VK_VERSION_1_1[] * *Scope* for *Non Uniform Group Operations* must: be limited to: ** *Subgroup* endif::VK_VERSION_1_1[] * *Storage Class* must: be limited to: ** *UniformConstant* ** *Input* ** *Uniform* ** *Output* ** *Workgroup* ** *Private* ** *Function* ** *PushConstant* ** *Image* ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ** *StorageBuffer* endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ifdef::VK_NV_ray_tracing[] ** *RayPayloadNV* ** *IncomingRayPayloadNV* ** *HitAttributeNV* ** *CallableDataNV* ** *IncomingCallableDataNV* ** *ShaderRecordBufferNV* endif::VK_NV_ray_tracing[] ifdef::VK_EXT_buffer_device_address[] ** *PhysicalStorageBufferEXT* endif::VK_EXT_buffer_device_address[] * Memory semantics must: obey the following rules: ** *Acquire* must: not be used with code:OpAtomicStore. ** *Release* must: not be used with code:OpAtomicLoad. ** *AcquireRelease* must: not be used with code:OpAtomicStore or code:OpAtomicLoad. ** Sequentially consistent atomics and barriers are not supported and *SequentiallyConsistent* is treated as *AcquireRelease*. *SequentiallyConsistent* should: not be used. ** code:OpMemoryBarrier must: use one of *Acquire*, *Release*, *AcquireRelease*, or *SequentiallyConsistent* and must: include at least one storage class. ** If the semantics for code:OpControlBarrier includes one of *Acquire*, *Release*, *AcquireRelease*, or *SequentiallyConsistent*, then it must: include at least one storage class. ** *SubgroupMemory*, *CrossWorkgroupMemory*, and *AtomicCounterMemory* are ignored. ifdef::VK_KHR_shader_float16_int8+!VK_KHR_8bit_storage[] * Any code:OpVariable with the result type pointing to an 8-bit integer object or an object containing an 8-bit integer element must: not have one of the following as its code:Storage code:Class operand: ** *Uniform* ** *PushConstant* ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ** *StorageBuffer* endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] endif::VK_KHR_shader_float16_int8+!VK_KHR_8bit_storage[] ifdef::VK_KHR_shader_float16_int8+!VK_KHR_16bit_storage[] * Any code:OpVariable with the result type pointing to a 16-bit floating-point object or an object containing a 16-bit floating-point element must: not have one of the following as its code:Storage code:Class operand: ** *Uniform* ** *PushConstant* ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ** *StorageBuffer* endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] endif::VK_KHR_shader_float16_int8+!VK_KHR_16bit_storage[] * Any code:OpVariable with an code:Initializer operand must: have one of the following as its *Storage Class* operand: ** *Output* ** *Private* ** *Function* * The code:OriginLowerLeft execution mode must: not be used; fragment entry points must: declare code:OriginUpperLeft. * The code:PixelCenterInteger execution mode must: not be used. Pixels are always centered at half-integer coordinates. * Images and Samplers ** code:OpTypeImage must: declare a scalar 32-bit float or 32-bit integer type for the "`Sampled Type`". (code:RelaxedPrecision can: be applied to a sampling instruction and to the variable holding the result of a sampling instruction.) ** code:OpTypeImage must: have a "`Sampled`" operand of 1 (sampled image) or 2 (storage image). ** If <> is not enabled and an code:OpTypeImage has "`Image Format`" operand of code:Unknown, any variables created with the given type must be decorated with code:NonReadable. ** If <> is not enabled and an code:OpTypeImage has "`Image Format`" operand of code:Unknown, any variables created with the given type must be decorated with code:NonWritable. ** code:OpImageQuerySizeLod, and code:OpImageQueryLevels must: only consume an "`Image`" operand whose type has its "`Sampled`" operand set to 1. ** The [eq]#(u,v)# coordinates used for a code:SubpassData must: be the of a constant vector [eq]#(0,0)#, or if a layer coordinate is used, must: be a vector that was formed with constant 0 for the [eq]#u# and [eq]#v# components. ** The "`Depth`" operand of code:OpTypeImage is ignored. ** Objects of types code:OpTypeImage, code:OpTypeSampler, code:OpTypeSampledImage, and arrays of these types must: not be stored to or modified. * Decorations ** Any code:BuiltIn decoration not listed in <> must: not be used. ** Any code:BuiltIn decoration that corresponds only to Vulkan features or extensions that have not been enabled must: not be used. ** The code:GLSLShared and code:GLSLPacked decorations must: not be used. ** The code:Flat, code:NoPerspective, code:Sample, and code:Centroid decorations must: not be used on variables with storage class other than code:Input or on variables used in the interface of non-fragment shader entry points. ** The code:Patch decoration must: not be used on variables in the interface of a vertex, geometry, or fragment shader stage's entry point. ifdef::VK_NV_viewport_array2[] ** The code:ViewportRelativeNV decoration must: only be used on a variable decorated with code:Layer in the vertex, tessellation evaluation, or geometry shader stages. ** The code:ViewportRelativeNV decoration must: not be used unless a variable decorated with one of code:ViewportIndex or code:ViewportMaskNV is also statically used by the same code:OpEntryPoint. ** The code:ViewportMaskNV and code:ViewportIndex decorations must: not both be statically used by one or more code:OpEntryPoint's that form the vertex processing stages of a graphics pipeline. endif::VK_NV_viewport_array2[] ifdef::VK_VERSION_1_1,VK_KHR_16bit_storage[] ** Only the round-to-nearest-even and the round-to-zero rounding modes can: be used for the code:FPRoundingMode decoration. ** The code:FPRoundingMode decoration can: only be used for the floating-point conversion instructions as described in the https://www.khronos.org/registry/spir-v/extensions/KHR/SPV_KHR_16bit_storage.html[`SPV_KHR_16bit_storage`] SPIR-V extension. endif::VK_VERSION_1_1,VK_KHR_16bit_storage[] ** code:DescriptorSet and code:Binding decorations must: obey the constraints on storage class, type, and descriptor type described in <> * code:OpTypeRuntimeArray must: only be used for: ** the last member of an code:OpTypeStruct ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] that is in the code:StorageBuffer storage class decorated as code:Block, or endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ifdef::VK_EXT_buffer_device_address[] that is in the code:PhysicalStorageBufferEXT storage class decorated as code:Block, or endif::VK_EXT_buffer_device_address[] that is in the code:Uniform storage class decorated as code:BufferBlock. ifdef::VK_EXT_descriptor_indexing[] ** If the code:RuntimeDescriptorArrayEXT capability is supported, an array of variables with storage class code:Uniform, ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] code:StorageBuffer, endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] or code:UniformConstant, or for the outermost dimension of an array of arrays of such variables. endif::VK_EXT_descriptor_indexing[] * Linkage: See <> for additional linking and validation rules. ifdef::VK_VERSION_1_1[] * If code:OpControlBarrier is used in fragment, vertex, tessellation evaluation, or geometry stages, the execution Scope must: be code:Subgroup. endif::VK_VERSION_1_1[] * Compute Shaders ** For each compute shader entry point, either a code:LocalSize execution mode or an object decorated with the code:WorkgroupSize decoration must: be specified. ifdef::VK_NV_compute_shader_derivatives[] ** For compute shaders using the code:DerivativeGroupQuadsNV execution mode, the first two dimensions of the local workgroup size must: be a multiple of two. ** 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. endif::VK_NV_compute_shader_derivatives[] ifdef::VK_VERSION_1_1[] * "`Result Type`" for *Non Uniform Group Operations* must: be limited to 32-bit float, 32-bit integer, boolean, or vectors of these types. If the code:Float64 capability is enabled, double and vectors of double types are also permitted. * If code:OpGroupNonUniformBallotBitCount is used, the group operation must: be one of: ** *Reduce* ** *InclusiveScan* ** *ExclusiveScan* endif::VK_VERSION_1_1[] * Atomic instructions must: declare a scalar 32-bit integer type, ifdef::VK_KHR_shader_atomic_int64[] or a scalar 64-bit integer type if the code:Int64Atomics capability is enabled, endif::VK_KHR_shader_atomic_int64[] for the value pointed to by _Pointer_. ifdef::VK_KHR_shader_atomic_int64[] ** <> must: be enabled for 64-bit integer atomic operations to be supported on a _Pointer_ with a *Storage Class* of ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] *StorageBuffer* or endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] *Uniform*. ** <> must: be enabled for 64-bit integer atomic operations to be supported on a _Pointer_ with a *Storage Class* of *Workgroup*. endif::VK_KHR_shader_atomic_int64[] * The _Pointer_ operand of all atomic instructions must: have a *Storage Class* limited to: ** *Uniform* ** *Workgroup* ** *Image* ifdef::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ** *StorageBuffer* endif::VK_VERSION_1_1,VK_KHR_storage_buffer_storage_class[] ifdef::VK_EXT_descriptor_indexing[] * If an instruction loads from or stores to a resource (including atomics and image instructions) and the resource descriptor being accessed is not dynamically uniform, then the operand corresponding to that resource (e.g. the pointer or sampled image operand) must: be decorated with code:NonUniformEXT. endif::VK_EXT_descriptor_indexing[] ifdef::VK_KHR_shader_float_controls[] * If <> is ename:VK_FALSE, then the entry point must: use the same denormals execution mode for both 16-bit and 64-bit floating-point types. * If <> is ename:VK_FALSE, then the entry point must: use the same rounding execution mode for both 16-bit and 64-bit floating-point types. * If <> is ename:VK_FALSE, then code:SignedZeroInfNanPreserve for 16-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:SignedZeroInfNanPreserve for 32-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:SignedZeroInfNanPreserve for 64-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormPreserve for 16-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormPreserve for 32-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormPreserve for 64-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormFlushToZero for 16-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormFlushToZero for 32-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:DenormFlushToZero for 64-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTE for 16-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTE for 32-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTE for 64-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTZ for 16-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTZ for 32-bit floating-point type must: not be used. * If <> is ename:VK_FALSE, then code:RoundingModeRTZ for 64-bit floating-point type must: not be used. endif::VK_KHR_shader_float_controls[] ifdef::VK_EXT_transform_feedback[] * The code:Offset plus size of the type of each variable, in the output interface of the entry point being compiled, decorated with code:XfbBuffer must: not be greater than sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:maxTransformFeedbackBufferDataSize * For any given code:XfbBuffer value, define the buffer data size to be smallest number of bytes such that, for all outputs decorated with the same code:XfbBuffer value, the size of the output interface variable plus the code:Offset is less than or equal to the buffer data size. For a given code:Stream, the sum of all the buffer data sizes for all buffers writing to that stream the must: not exceed sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:maxTransformFeedbackStreamDataSize * Output variables or block members decorated with code:Offset that have a 64-bit type, or a composite type containing a 64-bit type, must: specify an code:Offset value aligned to a 8 byte boundary * Any output block or block member decorated with code:Offset containing a 64-bit type consumes a multiple of 8 bytes * The size of any output block, that contains any member decorated with code:Offset that is a 64-bit type, must: be a multiple of 8 * The first member of an output block that specifies a code:Offset decoration must: specify a code:Offset value that is aligned to an 8 byte boundary if that block contains any member decorated with code:Offset and is a 64-bit type * Output variables or block members decorated with code:Offset that have a 32-bit type, or a composite type contains a 32-bit type, must: specify an code:Offset value aligned to a 4 byte boundary * Output variables, blocks or block members decorated with code:Offset must: only contain base types that have components that are either 32-bit or 64-bit in size * The Stream value to code:OpEmitStreamVertex and code:OpEndStreamPrimitive must: be less than sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:maxTransformFeedbackStreams * If the geometry shader emits to more than one vertex stream and sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:transformFeedbackStreamsLinesTriangles is ename:VK_FALSE, then execution mode must: be code:OutputPoints * Only variables or block members in the output interface decorated with code:Offset can: be captured for transform feedback, and those variables or block memebers must: also be decorated with code:XfbBuffer and code:XfbStride, or inherit code:XfbBuffer and code:XfbStride decorations from a block that contains them * All variables or block members in the output interface of the entry point being compiled decorated with a specific code:XfbBuffer value must: all be decorated with identical code:XfbStride values * If any variables or block members in the output interface of the entry point being compiled are decorated with code:Stream, then all variables belonging to the same code:XfbBuffer must specify the same code:Stream value * Output variables, blocks or block members that are not decorated with code:Stream default to vertex stream zero * For any two variables or block members in the output interface of the entry point being compiled with the same code:XfbBuffer value, the ranges determined by the code:Offset decoration and the size of the type must: not overlap * The stream number value to code:Stream must: be less than sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:maxTransformFeedbackStreams * The XFB Stride value to code:XfbStride must be less than or equal to sname:VkPhysicalDeviceTransformFeedbackPropertiesEXT::pname:maxTransformFeedbackBufferDataStride endif::VK_EXT_transform_feedback[] ifdef::VK_NV_ray_tracing[] * code:RayPayloadNV storage class must: only be used in ray generation, any-hit, closest hit or miss shaders. * code:IncomingRayPayloadNV storage class must: only be used in closest hit, any-hit, or miss shaders. * code:HitAttributeNV storage class must: only be used in intersection, any-hit, or closest hit shaders. * code:CallableDataNV storage class must: only be used in ray generation, closest hit, miss, and callable shaders. * code:IncomingCallableDataNV storage class must only be used in callable shaders. endif::VK_NV_ray_tracing[] ifdef::VK_VERSION_1_1,VK_KHR_variable_pointers,VK_EXT_buffer_device_address[] * The code:Base operand of code:OpPtrAccessChain must: point to one of the following storage classes: ifdef::VK_VERSION_1_1,VK_KHR_variable_pointers[] ** *Workgroup*, if code:VariablePointers is enabled. ** *StorageBuffer*, if code:VariablePointers or code:VariablePointersStorageBuffer is enabled. endif::VK_VERSION_1_1,VK_KHR_variable_pointers[] ** *PhysicalStorageBufferEXT*, if the code:PhysicalStorageBuffer64EXT addressing model is enabled. endif::VK_VERSION_1_1,VK_KHR_variable_pointers,VK_EXT_buffer_device_address[] ifdef::VK_EXT_buffer_device_address[] * If the code:PhysicalStorageBuffer64EXT addressing model is enabled: ** Any load or store through a physical pointer type must: be aligned to a multiple of the size of the largest scalar type in the pointed-to type. ** All instructions that support memory access operands and that use a physical pointer must: include the code:Aligned operand. ** The pointer value of a memory access instruction must be at least as aligned as specified by the code:Aligned memory access operand. ** Any access chain instruction that accesses into a code:RowMajor matrix must: only be used as the code:Pointer operand to code:OpLoad or code:OpStore. ** code:OpConvertUToPtr and code:OpConvertPtrToU must: use an integer type whose code:Width is 64. endif::VK_EXT_buffer_device_address[] ifdef::VK_NV_cooperative_matrix[] * For code:OpCooperativeMatrixLoadNV and code:OpCooperativeMatrixStoreNV instructions, the code:Pointer and code:Stride operands must: be aligned to at least the lesser of 16 bytes or the natural alignment of a row or column (depending on code:ColumnMajor) of the matrix (where the natural alignment is the number of columns/rows multiplied by the component size). * For code:OpTypeCooperativeMatrixNV, the component type, scope, number of rows, and number of columns must: match one of the matrices in any of the supported slink:VkCooperativeMatrixPropertiesNV. * For code:OpCooperativeMatrixMulAddNV, the code:Result, code:A, code:B, and code:C matrices must: all have types that satisfy the same supported slink:VkCooperativeMatrixPropertiesNV. That is, for one supported supported slink:VkCooperativeMatrixPropertiesNV, all of the following must: hold: ** The type of code:A must have pname:MSize rows and pname:KSize columns and have a component type that matches pname:AType. ** The type of code:B must have pname:KSize rows and pname:NSize columns and have a component type that matches pname:BType. ** The type of code:C must have pname:MSize rows and pname:NSize columns and have a component type that matches pname:CType. ** The type of code:Result must have pname:MSize rows and pname:NSize columns and have a component type that matches pname:DType. ** The type of code:A, code:B, code:C, and code:Result must all have a scope of pname:scope. * code:OpTypeCooperativeMatrixNV and code:OpCooperativeMatrix* instructions must: not be used in shader stages not included in slink:VkPhysicalDeviceCooperativeMatrixPropertiesNV::pname:cooperativeMatrixSupportedStages. endif::VK_NV_cooperative_matrix[] [[spirvenv-precision-operation]] == Precision and Operation of SPIR-V Instructions ifndef::VK_KHR_shader_float_controls[] The following rules apply to both single and double-precision floating point instructions: * Positive and negative infinities and positive and negative zeros are generated as dictated by <>, but subject to the precisions allowed in the following table. * Dividing a non-zero by a zero results in the appropriately signed <> infinity. * Any denormalized value input into a shader or potentially generated by any instruction in a shader may: be flushed to 0. * The rounding mode cannot: be set, and results will be <>, as described below. * [eq]##NaN##s may: not be generated. Instructions that operate on a [eq]#NaN# may: not result in a [eq]#NaN#. * Support for signaling [eq]##NaN##s is optional: and exceptions are never raised. endif::VK_KHR_shader_float_controls[] ifdef::VK_KHR_shader_float_controls[] The following rules apply to half, single, and double-precision floating point instructions: * Positive and negative infinities and positive and negative zeros are generated as dictated by <>, but subject to the precisions allowed in the following table. * Dividing a non-zero by a zero results in the appropriately signed <> infinity. * Signaling [eq]##NaN##s are not required to be generated and exceptions are never raised. Signaling [eq]##NaN## may: be converted to quiet [eq]##NaN##s values by any floating point instruction. * By default, the implementation may: perform optimizations on half, single, or double-precision floating-point instructions respectively that ignore sign of a zero, or assume that arguments and results are not [eq]##Nan##s or latexmath:[\pm\infty], this doesn't apply to code:OpIsNan and code:OpIsInf, which must: always correctly detect [eq]##Nan##s and latexmath:[\pm\infty]. If the entry point is declared with the code:SignedZeroInfNanPreserve execution mode, then sign of a zero, [eq]##Nan##s, and latexmath:[\pm\infty] must: not be ignored. ** The following core SPIR-V instructions must: respect the code:SignedZeroInfNanPreserve execution mode: code:OpPhi, code:OpSelect, code:OpReturnValue, code:OpVectorExtractDynamic, code:OpVectorInsertDynamic, code:OpVectorShuffle, code:OpCompositeConstruct, code:OpCompositeExtract, code:OpCompositeInsert, code:OpCopyObject, code:OpTranspose, code:OpFConvert, code:OpFNegate, code:OpFAdd, code:OpFSub, code:OpFMul, code:OpStore. This execution mode must: also be respected by code:OpLoad except for loads from the code:Input storage class in the fragment shader stage with the floating-point result type. Other SPIR-V instruction may: also respect the code:SignedZeroInfNanPreserve execution mode. * Denormalized values are supported. ** By default, any half, single, or double-precision denormalized value input into a shader or potentially generated by any instruction or any extended instructions for GLSL in a shader may: be flushed to zero. ** If the entry point is declared with the code:DenormFlushToZero execution mode then for the affected instuctions the denormalized result must: be flushed to zero and the denormalized operands may: be flushed to zero. Denormalized values obtained via unpacking an integer into a vector of values with smaller bit width and interpreting those values as floating-point numbers must: be flushed to zero. ** The following core SPIR-V instructions must: respect the code:DenormFlushToZero execution mode: code:OpSpecConstantOp (except when the opcode is code:OpQuantizeToF16), code:OpFConvert, code:OpFNegate, code:OpFAdd, code:OpFSub, code:OpFMul, code:OpFDiv, code:OpFRem, code:OpFMod, code:OpVectorTimesScalar, code:OpMatrixTimesScalar, code:OpVectorTimesMatrix, code:OpMatrixTimesVector, code:OpMatrixTimesMatrix, code:OpOuterProduct, code:OpDot; and the following extended instructions for GLSL: code:Round, code:RoundEven, code:Trunc, code:FAbs, code:Floor, code:Ceil, code:Fract, code:Radians, code:Degrees, code:Sin, code:Cos, code:Tan, code:Asin, code:Acos, code:Atan, code:Sinh, code:Cosh, code:Tanh, code:Asinh, code:Acosh, code:Atanh, code:Atan2, code:Pow, code:Exp, code:Log, code:Exp2, code:Log2, code:Sqrt, code:InverseSqrt, code:Determinant, code:MatrixInverse, code:Modf, code:ModfStruct, code:FMin, code:FMax, code:FClamp, code:FMix, code:Step, code:SmoothStep, code:Fma, code:UnpackHalf2x16, code:UnpackDouble2x32, code:Length, code:Distance, code:Cross, code:Normalize, code:FaceForward, code:Reflect, code:Refract, code:NMin, code:NMax, code:NClamp. Other SPIR-V instruction may: also respect the code:DenormFlushToZero execution mode. ** The following core SPIR-V instructions must: respect the code:DenormPreserve execution mode: code:OpPhi, code:OpSelect, code:OpReturnValue, code:OpVectorExtractDynamic, code:OpVectorInsertDynamic, code:OpVectorShuffle, code:OpCompositeConstruct, code:OpCompositeExtract, code:OpCompositeInsert, code:OpCopyObject, code:OpTranspose, code:OpStore, code:OpSpecConstantOp, code:OpFConvert, code:OpFNegate, code:OpFAdd, code:OpFSub, code:OpFMul, code:OpVectorTimesScalar, code:OpMatrixTimesScalar, code:OpVectorTimesMatrix, code:OpMatrixTimesVector, code:OpMatrixTimesMatrix, code:OpOuterProduct, code:OpDot, code:OpFOrdEqual, code:OpFUnordEqual, code:OpFOrdNotEqual, code:OpFUnordNotEqual, code:OpFOrdLessThan, code:OpFUnordLessThan, code:OpFOrdGreaterThan, code:OpFUnordGreaterThan, code:OpFOrdLessThanEqual, code:OpFUnordLessThanEqual, code:OpFOrdGreaterThanEqual, code:OpFUnordGreaterThanEqual; and the following extended instructions for GLSL: code:FAbs, code:FSign, code:Radians, code:Degrees, code:FMin, code:FMax, code:FClamp, code:FMix, code:Fma, code:PackHalf2x16, code:PackDouble2x32, code:UnpackHalf2x16, code:UnpackDouble2x32, code:NMin, code:NMax, code:NClamp. This execution mode must: also be respected by code:OpLoad except for loads from the code:Input storage class in the fragment shader stage with the floating-point result type. Other SPIR-V instruction may: also respect the code:DenormPreserve execution mode. endif::VK_KHR_shader_float_controls[] The precision of double-precision instructions is at least that of single precision. The precision of operations is defined either in terms of rounding, as an error bound in ULP, or as inherited from a formula as follows. .Correctly Rounded Operations described as "`correctly rounded`" will return the infinitely precise result, [eq]#x#, rounded so as to be representable in floating-point. ifdef::VK_KHR_shader_float_controls[] The rounding mode is not specified, unless the entry point is declared with the code:RoundingModeRTE or the code:RoundingModeRTZ execution mode. These execution modes affect only correctly rounded SPIR-V instructions. These execution modes do not affect code:OpQuantizeToF16. If the rounding mode is not specified then this rounding is implementation specific, subject to the following rules. endif::VK_KHR_shader_float_controls[] ifndef::VK_KHR_shader_float_controls[] The rounding mode used is not defined but must: obey the following rules. endif::VK_KHR_shader_float_controls[] If [eq]#x# is exactly representable then [eq]#x# will be returned. Otherwise, either the floating-point value closest to and no less than [eq]#x# or the value closest to and no greater than [eq]#x# will be returned. .ULP Where an error bound of [eq]#n# ULP (units in the last place) is given, for an operation with infinitely precise result #x# the value returned must: be in the range #[x - n * ulp(x), x + n * ulp(x)]#. The function #ulp(x)# is defined as follows: :: If there exist non-equal floating-point numbers #a# and #b# such that [eq]#a {leq} x {leq} b# then #ulp(x)# is the minimum possible distance between such numbers, latexmath:[ulp(x) = \mathrm{min}_{a,b} | b - a |]. If such numbers do not exist then #ulp(x)# is defined to be the difference between the two finite floating-point numbers nearest to #x#. Where the range of allowed return values includes any value of magnitude larger than that of the largest representable finite floating-point number, operations may:, additionally, return either an infinity of the appropriate sign or the finite number with the largest magnitude of the appropriate sign. If the infinitely precise result of the operation is not mathematically defined then the value returned is undefined:. .Inherited From ... Where an operation's precision is described as being inherited from a formula, the result returned must be at least as accurate as the result of computing an approximation to [eq]#x# using a formula equivalent to the given formula applied to the supplied inputs. Specifically, the formula given may be transformed using the mathematical associativity, commutativity and distributivity of the operators involved to yield an equivalent formula. The SPIR-V precision rules, when applied to each such formula and the given input values, define a range of permitted values. If [eq]#NaN# is one of the permitted values then the operation may return any result, otherwise let the largest permitted value in any of the ranges be [eq]#F~max~# and the smallest be [eq]#F~min~#. The operation must return a value in the range [eq]#[x - E, x + E]# where latexmath:[E = \mathrm{max} \left( | x - F_{\mathrm{min}} |, | x - F_{\mathrm{max}} | \right) ]. ifdef::VK_KHR_shader_float_controls[] If the entry point is declared with the code:DenormFlushToZero execution mode, then any intermediate denormal value(s) while evaluating the formula may: be flushed to zero. Denormal final results must: be flushed to zero. If the entry point is declared with the code:DenormPreserve execution mode, then denormals must: be preserved throughout the formula. endif::VK_KHR_shader_float_controls[] ifdef::VK_KHR_shader_float16_int8[] For half- (16 bit) and single- (32 bit) precision instructions, precisions are required: to be at least as follows: .Precision of core SPIR-V Instructions [options="header", cols=",,"] |==== | Instruction | Single precision, unless decorated with RelaxedPrecision | Half precision | code:OpFAdd 2+| Correctly rounded. | code:OpFSub 2+| Correctly rounded. | code:OpFMul, code:OpVectorTimesScalar, code:OpMatrixTimesScalar 2+| Correctly rounded. | code:OpDot(x, y) 2+| Inherited from latexmath:[\sum_{i = 0}^{n - 1} x_{i} \times y_{i}]. | code:OpFOrdEqual, code:OpFUnordEqual 2+| Correct result. | code:OpFOrdLessThan, code:OpFUnordLessThan 2+| Correct result. | code:OpFOrdGreaterThan, code:OpFUnordGreaterThan 2+| Correct result. | code:OpFOrdLessThanEqual, code:OpFUnordLessThanEqual 2+| Correct result. | code:OpFOrdGreaterThanEqual, code:OpFUnordGreaterThanEqual 2+| Correct result. | code:OpFDiv(x,y) | 2.5 ULP for y in the range [2^-126^, 2^126^]. | 2.5 ULP for y in the range [2^-14^, 2^14^]. | code:OpFRem(x,y) | Inherited from [eq]#x - y {times} trunc(x/y)#, for y in the range [2^-126^, 2^126^]. | Inherited from [eq]#x - y {times} trunc(x/y)#, for y in the range [2^-14^, 2^14^]. | code:OpFMod(x,y) | Inherited from [eq]#x - y {times} floor(x/y)#, for y in the range [2^-126^, 2^126^]. | Inherited from [eq]#x - y {times} floor(x/y)#, for y in the range [2^-14^, 2^14^]. | conversions between types 2+| Correctly rounded. |==== [NOTE] .Note ==== The code:OpFRem and code:OpFMod instructions use cheap approximations of remainder, and the error can be large due to the discontinuity in trunc() and floor(). This can produce mathematically unexpected results in some cases, such as FMod(x,x) computing x rather than 0, and can also cause the result to have a different sign than the infinitely precise result. ==== .Precision of GLSL.std.450 Instructions [options="header", cols=",,"] |==== |Instruction | Single precision, unless decorated with RelaxedPrecision | Half precision | code:fma() 2+| Inherited from code:OpFMul followed by code:OpFAdd. | code:exp(x), code:exp2(x) | [eq]#3 + 2 {times} {vert}x{vert}# ULP. | [eq]#1 + 2 {times} {vert}x{vert}# ULP. | code:log(), code:log2() | 3 ULP outside the range [eq]#[0.5, 2.0]#. Absolute error < [eq]#2^-21^# inside the range [eq]#[0.5, 2.0]#. | 3 ULP outside the range [eq]#[0.5, 2.0]#. Absolute error < [eq]#2^-7^# inside the range [eq]#[0.5, 2.0]#. | code:pow(x, y) 2+| Inherited from code:exp2(y {times} code:log2(x)). | code:sqrt() 2+| Inherited from 1.0 / code:inversesqrt(). | code:inversesqrt() 2+| 2 ULP. | code:radians(x) 2+| Inherited from latexmath:[\frac{x \times \pi}{180}]. | code:degrees(x) 2+| Inherited from latexmath:[\frac{x \times 180}{\pi}]. | code:sin() | Absolute error latexmath:[\leq 2^{-11}] inside the range latexmath:[[-\pi, \pi\]]. | Absolute error latexmath:[\leq 2^{-7}] inside the range latexmath:[[-\pi, \pi\]]. | code:cos() | Absolute error latexmath:[\leq 2^{-11}] inside the range latexmath:[[-\pi, \pi\]]. | Absolute error latexmath:[\leq 2^{-7}] inside the range latexmath:[[-\pi, \pi\]]. | code:tan() 2+| Inherited from latexmath:[\frac{sin()}{cos()}]. | code:asin(x) 2+| Inherited from latexmath:[atan2(x, sqrt(1.0 - x^2))]. | code:acos(x) 2+| Inherited from latexmath:[atan2(sqrt(1.0 - x^2), x)]. | code:atan(), code:atan2() | 4096 ULP | 5 ULP. | code:sinh(x) 2+| Inherited from latexmath:[(exp(x) - exp(-x)) \times 0.5]. | code:cosh(x) 2+| Inherited from latexmath:[(exp(x) + exp(-x)) \times 0.5]. | code:tanh() 2+| Inherited from latexmath:[\frac{sinh()}{cosh()}]. | code:asinh(x) 2+| Inherited from latexmath:[log(x + sqrt(x^2 + 1.0))]. | code:acosh(x) 2+| Inherited from latexmath:[log(x + sqrt(x^2 - 1.0))]. | code:atanh(x) 2+| Inherited from latexmath:[log(\frac{1.0 + x}{1.0 - x}) \times 0.5]. | code:frexp() 2+| Correctly rounded. | code:ldexp() 2+| Correctly rounded. | code:length(x) 2+| Inherited from latexmath:[sqrt(dot(x, x))]. | code:distance(x, y) 2+| Inherited from latexmath:[length(x - y)]. | code:cross() 2+| Inherited from [eq]#code:OpFSub(code:OpFMul, code:OpFMul)#. | code:normalize(x) 2+| Inherited from latexmath:[\frac{x}{length(x)}]. | code:faceforward 2+| Correctly rounded. | code:reflect(x, y) 2+| Inherited from [eq]#x - 2.0 {times} code:dot(y, x) {times} y#. | code:refract(I, N, eta) 2+| Inherited from [eq]#eta {times} I - (eta {times} code:dot(N, I) + code:sqrt(k)) {times} N#. | code:round 2+| Correctly rounded. | code:roundEven 2+| Correctly rounded. | code:trunc 2+| Correctly rounded. | code:fabs 2+| Correctly rounded. | code:fsign 2+| Correctly rounded. | code:floor 2+| Correctly rounded. | code:ceil 2+| Correctly rounded. | code:fract 2+| Correctly rounded. | code:modf 2+| Correctly rounded. | code:fmin 2+| Correctly rounded. | code:fmax 2+| Correctly rounded. | code:fclamp 2+| Correctly rounded. | code:fmix(x, y, a) 2+| Inherited from [eq]#x {times} (1.0 - a) + y {times} a#. | code:step 2+| Correctly rounded. | code:smoothStep(edge0, edge1, x) 2+| Inherited from [eq]#t {times} t {times} (3.0 - 2.0 {times} t)#, where latexmath:[t = clamp(\frac{x - edge0}{edge1 - edge0}, 0.0, 1.0)]. | code:nmin 2+| Correctly rounded. | code:nmax 2+| Correctly rounded. | code:nclamp 2+| Correctly rounded. |==== endif::VK_KHR_shader_float16_int8[] ifndef::VK_KHR_shader_float16_int8[] For single precision (32 bit) instructions, precisions are required: to be at least as follows, unless decorated with RelaxedPrecision: .Precision of core SPIR-V Instructions [options="header"] |==== | Instruction | Precision | code:OpFAdd | Correctly rounded. | code:OpFSub | Correctly rounded. | code:OpFMul, code:OpVectorTimesScalar, code:OpMatrixTimesScalar | Correctly rounded. | code:OpFOrdEqual, code:OpFUnordEqual | Correct result. | code:OpFOrdLessThan, code:OpFUnordLessThan | Correct result. | code:OpFOrdGreaterThan, code:OpFUnordGreaterThan | Correct result. | code:OpFOrdLessThanEqual, code:OpFUnordLessThanEqual | Correct result. | code:OpFOrdGreaterThanEqual, code:OpFUnordGreaterThanEqual | Correct result. | code:OpFDiv | 2.5 ULP for b in the range [2^-126^, 2^126^]. | conversions between types | Correctly rounded. |==== .Precision of GLSL.std.450 Instructions [options="header"] |==== |Instruction | Precision | code:fma() | Inherited from code:OpFMul followed by code:OpFAdd. | code:exp(x), code:exp2(x) | [eq]#3 {plus} 2 {times} {vert}x{vert}# ULP. | code:log(), code:log2() | 3 ULP outside the range [eq]#[0.5, 2.0]#. Absolute error < [eq]#2^-21^# inside the range [eq]#[0.5, 2.0]#. | code:pow(x, y) | Inherited from code:exp2(y {times} code:log2(x)). | code:sqrt() | Inherited from 1.0 / code:inversesqrt(). | code:inversesqrt() | 2 ULP. |==== endif::VK_KHR_shader_float16_int8[] GLSL.std.450 extended instructions specifically defined in terms of the above instructions inherit the above errors. GLSL.std.450 extended instructions not listed above and not defined in terms of the above have undefined: precision. These include, for example, the trigonometric functions and determinant. For the code:OpSRem and code:OpSMod instructions, if either operand is negative the result is undefined:. [NOTE] .Note ==== While the code:OpSRem and code:OpSMod instructions are supported by the Vulkan environment, they require non-negative values and thus do not enable additional functionality beyond what code:OpUMod provides. ==== ifdef::VK_NV_cooperative_matrix[] code:OpCooperativeMatrixMulAddNV performs its operations in an implementation-dependent order and internal precision. endif::VK_NV_cooperative_matrix[] [[spirvenv-image-formats]] == Compatibility Between SPIR-V Image Formats And Vulkan Formats Images which are read from or written to by shaders must: have SPIR-V image formats compatible with the Vulkan image formats backing the image under the circumstances described for <>. The compatibile formats are: .SPIR-V and Vulkan Image Format Compatibility [cols="2*", options="header"] |==== |SPIR-V Image Format |Compatible Vulkan Format |code:Rgba32f |ename:VK_FORMAT_R32G32B32A32_SFLOAT |code:Rgba16f |ename:VK_FORMAT_R16G16B16A16_SFLOAT |code:R32f |ename:VK_FORMAT_R32_SFLOAT |code:Rgba8 |ename:VK_FORMAT_R8G8B8A8_UNORM |code:Rgba8Snorm |ename:VK_FORMAT_R8G8B8A8_SNORM |code:Rg32f |ename:VK_FORMAT_R32G32_SFLOAT |code:Rg16f |ename:VK_FORMAT_R16G16_SFLOAT |code:R11fG11fB10f |ename:VK_FORMAT_B10G11R11_UFLOAT_PACK32 |code:R16f |ename:VK_FORMAT_R16_SFLOAT |code:Rgba16 |ename:VK_FORMAT_R16G16B16A16_UNORM |code:Rgb10A2 |ename:VK_FORMAT_A2B10G10R10_UNORM_PACK32 |code:Rg16 |ename:VK_FORMAT_R16G16_UNORM |code:Rg8 |ename:VK_FORMAT_R8G8_UNORM |code:R16 |ename:VK_FORMAT_R16_UNORM |code:R8 |ename:VK_FORMAT_R8_UNORM |code:Rgba16Snorm |ename:VK_FORMAT_R16G16B16A16_SNORM |code:Rg16Snorm |ename:VK_FORMAT_R16G16_SNORM |code:Rg8Snorm |ename:VK_FORMAT_R8G8_SNORM |code:R16Snorm |ename:VK_FORMAT_R16_SNORM |code:R8Snorm |ename:VK_FORMAT_R8_SNORM |code:Rgba32i |ename:VK_FORMAT_R32G32B32A32_SINT |code:Rgba16i |ename:VK_FORMAT_R16G16B16A16_SINT |code:Rgba8i |ename:VK_FORMAT_R8G8B8A8_SINT |code:R32i |ename:VK_FORMAT_R32_SINT |code:Rg32i |ename:VK_FORMAT_R32G32_SINT |code:Rg16i |ename:VK_FORMAT_R16G16_SINT |code:Rg8i |ename:VK_FORMAT_R8G8_SINT |code:R16i |ename:VK_FORMAT_R16_SINT |code:R8i |ename:VK_FORMAT_R8_SINT |code:Rgba32ui |ename:VK_FORMAT_R32G32B32A32_UINT |code:Rgba16ui |ename:VK_FORMAT_R16G16B16A16_UINT |code:Rgba8ui |ename:VK_FORMAT_R8G8B8A8_UINT |code:R32ui |ename:VK_FORMAT_R32_UINT |code:Rgb10a2ui |ename:VK_FORMAT_A2B10G10R10_UINT_PACK32 |code:Rg32ui |ename:VK_FORMAT_R32G32_UINT |code:Rg16ui |ename:VK_FORMAT_R16G16_UINT |code:Rg8ui |ename:VK_FORMAT_R8G8_UINT |code:R16ui |ename:VK_FORMAT_R16_UINT |code:R8ui |ename:VK_FORMAT_R8_UINT |====