Capability Atoms¶

Sections:¶

Targets
Stages
Versions
Extensions
Compound Capabilities
Other

Targets¶

Capabilities to specify code generation targets (glsl, spirv…)

c

Represents the C programming language code generation target.

cpp

Represents the C++ programming language code generation target.

cuda

Represents the CUDA code generation target.

glsl

Represents the GLSL code generation target.

hlsl

Represents the HLSL code generation target.

llvm

Represents the LLVM IR target.

metal

Represents the Metal programming language code generation target.

slangvm

Represents the Slang VM bytecode target.

spirv

Represents the SPIR-V code generation target.

textualTarget

Represents a non-assembly code generation target.

wgsl

Represents the WebGPU shading language code generation target.

Stages¶

Capabilities to specify code generation stages (vertex, fragment…)

amplification

Amplification shader stage & mesh shader capabilities

anyhit

Any-Hit shader stage & ray-tracing capabilities

callable

Callable shader stage & ray-tracing capabilities

closesthit

Closest-Hit shader stage & ray-tracing capabilities

compute

Compute shader stage

dispatch

Dispatch shader stage

domain

Domain shader stage

fragment

Fragment shader stage

geometry

Geometry shader stage

hull

Hull shader stage

intersection

Intersection shader stage & ray-tracing capabilities

mesh

Mesh shader stage & mesh shader capabilities

miss

Ray-Miss shader stage & ray-tracing capabilities

pixel

Pixel shader stage

raygen

Ray-Generation shader stage & ray-tracing capabilities

raygeneration

Ray-Generation shader stage & ray-tracing capabilities

task

Task shader stage & mesh shader capabilities

tesscontrol

Tessellation Control shader stage

tesseval

Tessellation Evaluation shader stage

vertex

Vertex shader stage

Versions¶

Capabilities to specify versions of a code generation target (sm_5_0, GLSL_400…)

GLSL_130

GLSL 130 and related capabilities of other targets.

GLSL_140

GLSL 140 and related capabilities of other targets.

GLSL_150

GLSL 150 and related capabilities of other targets.

GLSL_330

GLSL 330 and related capabilities of other targets.

GLSL_400

GLSL 400 and related capabilities of other targets.

GLSL_410

GLSL 410 and related capabilities of other targets.

GLSL_420

GLSL 420 and related capabilities of other targets.

GLSL_430

GLSL 430 and related capabilities of other targets.

GLSL_440

GLSL 440 and related capabilities of other targets.

GLSL_450

GLSL 450 and related capabilities of other targets.

GLSL_460

GLSL 460 and related capabilities of other targets.

cuda_sm_1_0

cuda 1.0 and related capabilities of other targets.

cuda_sm_2_0

cuda 2.0 and related capabilities of other targets.

cuda_sm_3_0

cuda 3.0 and related capabilities of other targets.

cuda_sm_3_5

cuda 3.5 and related capabilities of other targets.

cuda_sm_4_0

cuda 4.0 and related capabilities of other targets.

cuda_sm_5_0

cuda 5.0 and related capabilities of other targets.

cuda_sm_6_0

cuda 6.0 and related capabilities of other targets.

cuda_sm_7_0

cuda 7.0 and related capabilities of other targets.

cuda_sm_8_0

cuda 8.0 and related capabilities of other targets.

cuda_sm_8_9

cuda 8.9 (Ada Lovelace) and related capabilities of other targets. Required for PTX features that landed in PTX ISA 8.7 / SM 8.9, notably the mma.sync.m16n8k16 instruction with .e4m3 / .e5m2 floating-point inputs (FP8 cooperative matrices).

cuda_sm_9_0

cuda 9.0 and related capabilities of other targets.

dxil_lib

Represents capabilities required for DXIL Library compilation.

glsl_spirv_1_0

Represents SPIR-V 1.0 through glslang.

glsl_spirv_1_1

Represents SPIR-V 1.1 through glslang.

glsl_spirv_1_2

Represents SPIR-V 1.2 through glslang.

glsl_spirv_1_3

Represents SPIR-V 1.3 through glslang.

glsl_spirv_1_4

Represents SPIR-V 1.4 through glslang.

glsl_spirv_1_5

Represents SPIR-V 1.5 through glslang.

glsl_spirv_1_6

Represents SPIR-V 1.6 through glslang.

hlsl_2018

Represent HLSL compatibility support.

hlsl_nvapi

Represents HLSL NVAPI support.

metallib_2_3

Represents MetalLib 2.3.

metallib_2_4

Represents MetalLib 2.4.

metallib_3_0

Represents MetalLib 3.0.

metallib_3_1

Represents MetalLib 3.1.

metallib_4_0

Represents MetalLib 4.0.

metallib_latest

Represents the latest MetalLib version.

sm_4_0

HLSL shader model 4.0 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_4_0_version

HLSL shader model 4.0 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_4_1

HLSL shader model 4.1 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_4_1_version

HLSL shader model 4.1 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_5_0

HLSL shader model 5.0 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_5_0_version

HLSL shader model 5.0 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_5_1

HLSL shader model 5.1 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_5_1_version

HLSL shader model 5.1 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_0

HLSL shader model 6.0 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_0_version

HLSL shader model 6.0 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_1

HLSL shader model 6.1 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_10

HLSL shader model 6.10 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_10_version

HLSL shader model 6.10 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_1_version

HLSL shader model 6.1 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_2

HLSL shader model 6.2 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_2_version

HLSL shader model 6.2 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_3

HLSL shader model 6.3 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_3_version

HLSL shader model 6.3 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_4

HLSL shader model 6.4 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_4_version

HLSL shader model 6.4 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_5

HLSL shader model 6.5 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_5_version

HLSL shader model 6.5 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_6

HLSL shader model 6.6 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_6_version

HLSL shader model 6.6 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_7

HLSL shader model 6.7 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_7_version

HLSL shader model 6.7 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_8

HLSL shader model 6.8 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_8_version

HLSL shader model 6.8 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

sm_6_9

HLSL shader model 6.9 and related capabilities of other targets. Includes related GLSL/SPIRV extensions.

sm_6_9_version

HLSL shader model 6.9 and related capabilities of other targets. Does not include related GLSL/SPIRV extensions.

spirv_1_0

Represents SPIR-V 1.0 version.

spirv_1_1

Represents SPIR-V 1.1 version, which includes SPIR-V 1.0.

spirv_1_2

Represents SPIR-V 1.2 version, which includes SPIR-V 1.1.

spirv_1_3

Represents SPIR-V 1.3 version, which includes SPIR-V 1.2.

spirv_1_4

Represents SPIR-V 1.4 version, which includes SPIR-V 1.3.

spirv_1_5

Represents SPIR-V 1.5 version, which includes SPIR-V 1.4 and additional extensions.

spirv_1_6

Represents SPIR-V 1.6 version, which includes SPIR-V 1.5 and additional extensions.

spirv_latest

Represents the latest SPIR-V version.

Extensions¶

Capabilities to specify extensions (GL_EXT, SPV_EXT…)

GL_ARB_derivative_control

Represents the GL_ARB_derivative_control extension.

GL_ARB_fragment_shader_interlock

Represents the GL_ARB_fragment_shader_interlock extension.

GL_ARB_gpu_shader5

Represents the GL_ARB_gpu_shader5 extension.

GL_ARB_gpu_shader_int64

Represents the GL_ARB_gpu_shader_int64 extension.

GL_ARB_shader_clock

Represents the GL_ARB_shader_clock extension.

GL_ARB_shader_clock64

Represents the GL_ARB_shader_clock64 extension.

GL_ARB_shader_image_load_store

Represents the GL_ARB_shader_image_load_store extension.

GL_ARB_shader_image_size

Represents the GL_ARB_shader_image_size extension.

GL_ARB_shader_texture_image_samples

Represents the GL_ARB_shader_texture_image_samples extension.

GL_ARB_sparse_texture

Represents the GL_ARB_sparse_texture extension.

GL_ARB_sparse_texture2

Represents the GL_ARB_sparse_texture2 extension.

GL_ARB_sparse_texture_clamp

Represents the GL_ARB_sparse_texture_clamp extension.

GL_ARB_texture_gather

Represents the GL_ARB_texture_gather extension.

GL_ARB_texture_multisample

Represents the GL_ARB_texture_multisample extension.

GL_ARB_texture_query_levels

Represents the GL_ARB_texture_query_levels extension.

GL_EXT_buffer_reference

Represents the GL_EXT_buffer_reference extension.

GL_EXT_buffer_reference_uvec2

Represents the GL_EXT_buffer_reference_uvec2 extension.

GL_EXT_debug_printf

Represents the GL_EXT_debug_printf extension.

GL_EXT_demote_to_helper_invocation

Represents the GL_EXT_demote_to_helper_invocation extension.

GL_EXT_device_group

Represents the GL_EXT_device_group extension.

GL_EXT_fragment_shader_barycentric

Represents the GL_EXT_fragment_shader_barycentric extension.

GL_EXT_maximal_reconvergence

Represents the GL_EXT_maximal_reconvergence extension.

GL_EXT_mesh_shader

Represents the GL_EXT_mesh_shader extension.

GL_EXT_nonuniform_qualifier

Represents the GL_EXT_nonuniform_qualifier extension.

GL_EXT_ray_query

Represents the GL_EXT_ray_query extension.

GL_EXT_ray_tracing

Represents the GL_EXT_ray_tracing extension.

GL_EXT_ray_tracing_position_fetch

Represents the GL_EXT_ray_tracing_position_fetch extension.

GL_EXT_ray_tracing_position_fetch_ray_query

Represents the GL_EXT_ray_tracing_position_fetch_ray_query extension.

GL_EXT_ray_tracing_position_fetch_ray_tracing

Represents the GL_EXT_ray_tracing_position_fetch_ray_tracing extension.

GL_EXT_samplerless_texture_functions

Represents the GL_EXT_samplerless_texture_functions extension.

GL_EXT_shader_atomic_float

Represents the GL_EXT_shader_atomic_float extension.

GL_EXT_shader_atomic_float2

Represents the GL_EXT_shader_atomic_float2 extension.

GL_EXT_shader_atomic_float_min_max

Represents the GL_EXT_shader_atomic_float_min_max extension.

GL_EXT_shader_atomic_int64

Represents the GL_EXT_shader_atomic_int64 extension.

GL_EXT_shader_explicit_arithmetic_types

Represents the GL_EXT_shader_explicit_arithmetic_types extension.

GL_EXT_shader_explicit_arithmetic_types_int64

Represents the GL_EXT_shader_explicit_arithmetic_types_int64 extension.

GL_EXT_shader_image_load_store

Represents the GL_EXT_shader_image_load_store extension.

GL_EXT_shader_invocation_reorder

Represents the GL_EXT_shader_invocation_reorder extension (cross-vendor standard).

GL_EXT_shader_invocation_reorder_motion

Represents the GL_EXT_shader_invocation_reorder + motion blur combined extension (cross-vendor SER with NV motion blur). Used for motion blur variants of EXT SER functions.

GL_EXT_shader_quad_control

Represents the GL_EXT_shader_quad_control extension.

GL_EXT_shader_realtime_clock

Represents the GL_EXT_shader_realtime_clock extension.

GL_EXT_texture_query_lod

Represents the GL_EXT_texture_query_lod extension.

GL_EXT_texture_shadow_lod

Represents the GL_EXT_texture_shadow_lod extension.

GL_KHR_memory_scope_semantics

Represents the GL_KHR_memory_scope_semantics extension.

GL_KHR_shader_subgroup_arithmetic

Represents the GL_KHR_shader_subgroup_arithmetic extension.

GL_KHR_shader_subgroup_ballot

Represents the GL_KHR_shader_subgroup_ballot extension.

GL_KHR_shader_subgroup_basic

Represents the GL_KHR_shader_subgroup_basic extension.

GL_KHR_shader_subgroup_clustered

Represents the GL_KHR_shader_subgroup_clustered extension.

GL_KHR_shader_subgroup_quad

Represents the GL_KHR_shader_subgroup_quad extension.

GL_KHR_shader_subgroup_rotate

Represents the GL_KHR_shader_subgroup_rotate extension.

GL_KHR_shader_subgroup_shuffle

Represents the GL_KHR_shader_subgroup_shuffle extension.

GL_KHR_shader_subgroup_shuffle_relative

Represents the GL_KHR_shader_subgroup_shuffle_relative extension.

GL_KHR_shader_subgroup_vote

Represents the GL_KHR_shader_subgroup_vote extension.

GL_NV_cluster_acceleration_structure

Represents the GL_NV_cluster_acceleration_structure extension.

GL_NV_compute_shader_derivatives

Represents the GL_NV_compute_shader_derivatives extension.

This capability enables the use of implicit derivatives in compute, ray tracing, and mesh stages.

This capability changes the interpretation of GLSL implicit-LOD texture sampling functions as follows, matching the GLSL shader specification:

Derivatives enabled: Implicit-LOD texture() functions are assumed to use implicit LOD.

Derivatives disabled: Implicit-LOD texture() functions are assumed to use the base texture.

This applies to GLSL as both source and target.

GL_NV_cooperative_vector

Represents the GL_NV_cooperative_vector extension.

GL_NV_fragment_shader_barycentric

Represents the GL_NV_fragment_shader_barycentric extension.

GL_NV_gpu_shader5

Represents the GL_NV_gpu_shader5 extension.

GL_NV_ray_tracing

Represents the GL_NV_ray_tracing extension.

GL_NV_ray_tracing_motion_blur

Represents the GL_NV_ray_tracing_motion_blur extension.

GL_NV_shader_atomic_fp16_vector

Represents the GL_NV_shader_atomic_fp16_vector extension.

GL_NV_shader_invocation_reorder

Represents the GL_NV_shader_invocation_reorder extension (NVIDIA-specific).

GL_NV_shader_invocation_reorder_motion

Represents the GL_NV_shader_invocation_reorder + motion blur combined extension (NVIDIA-specific). Used for motion blur variants of SER functions that require both extensions.

GL_NV_shader_subgroup_partitioned

Represents the GL_NV_shader_subgroup_partitioned extension.

GL_NV_shader_texture_footprint

Represents the GL_NV_shader_texture_footprint extension.

SPV_EXT_demote_to_helper_invocation

Represents the SPIR-V extension for demoting to helper invocation.

SPV_EXT_descriptor_heap

Represents the SPIR-V extension for descriptor heaps.

SPV_EXT_descriptor_indexing

Represents the SPIR-V extension for descriptor indexing.

SPV_EXT_float8

Represents the SPIR-V extension for SPV_EXT_float8.

SPV_EXT_fragment_fully_covered

Represents the SPIR-V extension for SPV_EXT_fragment_fully_covered.

SPV_EXT_fragment_shader_interlock

Represents the SPIR-V extension for fragment shader interlock operations.

SPV_EXT_mesh_shader

Represents the SPIR-V extension for mesh shaders.

SPV_EXT_physical_storage_buffer

Represents the SPIR-V extension for physical storage buffer.

SPV_EXT_replicated_composites

Represents the SPIR-V extension for SPV_EXT_replicated_composites.

SPV_EXT_shader_atomic_float16_add

Represents the SPIR-V extension for atomic float16 add operations.

SPV_EXT_shader_atomic_float_add

Represents the SPIR-V extension for atomic float add operations.

SPV_EXT_shader_atomic_float_min_max

Represents the SPIR-V extension for atomic float min/max operations.

SPV_EXT_shader_invocation_reorder

Represents the SPIR-V extension for shader invocation reorder (cross-vendor standard). Requires SPV_KHR_ray_tracing and SPIR-V 1.5 (which includes physical storage buffer). Note: Spec allows SPIR-V 1.4 + physical_storage_buffer extension, but we require 1.5 for simplicity.

SPV_GOOGLE_user_type

Represents the SPIR-V extension for SPV_GOOGLE_user_type.

SPV_KHR_bfloat16

Represents the SPIR-V extension for BFloat16 types.

SPV_KHR_compute_shader_derivatives

Represents the SPIR-V extension for compute shader derivatives.

SPV_KHR_cooperative_matrix

Represents the SPIR-V extension for SPV_KHR_cooperative_matrix.

SPV_KHR_device_group

Represents the SPIR-V extension for device-group information.

SPV_KHR_fragment_shader_barycentric

Represents the SPIR-V extension for fragment shader barycentric.

SPV_KHR_maximal_reconvergence

Represents the SPIR-V extension for maximal reconvergence.

SPV_KHR_non_semantic_info

Represents the SPIR-V extension for non-semantic information.

SPV_KHR_quad_control

Represents the SPIR-V extension for quad group control.

SPV_KHR_ray_query

Represents the SPIR-V extension for ray queries.

SPV_KHR_ray_tracing

Represents the SPIR-V extension for ray tracing.

SPV_KHR_ray_tracing_position_fetch

Represents the SPIR-V extension for ray tracing position fetch. Should be used with either SPV_KHR_ray_query or SPV_KHR_ray_tracing.

SPV_KHR_shader_clock

Represents the SPIR-V extension for shader clock.

SPV_KHR_subgroup_rotate

Represents the SPIR-V extension enables rotating values across invocations within a subgroup.

SPV_KHR_untyped_pointers

Represents the SPIR-V extension for untyped pointers.

SPV_KHR_vulkan_memory_model

Represents the SPIR-V extension for SPV_KHR_vulkan_memory_model.

SPV_NV_bindless_texture

Represents the SPIR-V extension for SPV_NV_bindless_texture.

SPV_NV_cluster_acceleration_structure

Represents the SPIR-V extension for cluster acceleration structure.

SPV_NV_compute_shader_derivatives

Represents the SPIR-V extension for compute shader derivatives.

SPV_NV_cooperative_matrix2

Represents the SPIR-V extension for SPV_NV_cooperative_matrix2.

SPV_NV_cooperative_vector

Represents the SPIR-V extension for SPV_NV_cooperative_vector.

SPV_NV_linear_swept_spheres

Represents the SPIR-V extension for linear swept spheres.

SPV_NV_ray_tracing_motion_blur

Represents the SPIR-V extension for ray tracing motion blur.

SPV_NV_shader_image_footprint

Represents the SPIR-V extension for shader image footprint.

SPV_NV_shader_invocation_reorder

Represents the SPIR-V extension for shader invocation reorder (NVIDIA-specific). Inherits from SPV_EXT_shader_invocation_reorder so NV implies EXT.

SPV_NV_shader_subgroup_partitioned

Represents the SPIR-V extension for shader subgroup partitioned.

SPV_NV_tensor_addressing

Represents the SPIR-V extension for SPV_NV_tensor_addressing.

ser_hlsl_native

DXR 1.3 native SER support (SM 6.9, no NVAPI required)

spvAtomicFloat16AddEXT

Represents the SPIR-V capability for atomic float 16 add operations.

spvAtomicFloat16MinMaxEXT

Represents the SPIR-V capability for atomic float 16 min/max operations.

spvAtomicFloat32AddEXT

Represents the SPIR-V capability for atomic float 32 add operations.

spvAtomicFloat32MinMaxEXT

Represents the SPIR-V capability for atomic float 32 min/max operations.

spvAtomicFloat64AddEXT

Represents the SPIR-V capability for atomic float 64 add operations.

spvAtomicFloat64MinMaxEXT

Represents the SPIR-V capability for atomic float 64 min/max operations.

spvBFloat16KHR

Represents the SPIR-V capability for using bf16 floating point type.

spvBindlessTextureNV

Represents the SPIR-V capability for the bindless texture.

spvCooperativeMatrixBlockLoadsNV

Represents the SPIR-V capability for cooperative matrix 2

spvCooperativeMatrixConversionsNV

Represents the SPIR-V capability for cooperative matrix 2

spvCooperativeMatrixKHR

Represents the SPIR-V capability for cooperative matrices

spvCooperativeMatrixPerElementOperationsNV

Represents the SPIR-V capability for cooperative matrix 2

spvCooperativeMatrixReductionsNV

Represents the SPIR-V capability for cooperative matrix 2

spvCooperativeMatrixTensorAddressingNV

Represents the SPIR-V capability for cooperative matrix 2

spvCooperativeVectorNV

Represents the SPIR-V capability for cooperative vectors

spvCooperativeVectorTrainingNV

Represents the SPIR-V capability for cooperative vector training

spvDemoteToHelperInvocation

Represents the SPIR-V capability for demoting to helper invocation.

spvDemoteToHelperInvocationEXT

Represents the SPIR-V capability for demoting to helper invocation.

spvDerivativeControl

Represents the SPIR-V capability for ‘derivative control’ operations.

spvDescriptorHeapEXT

Represents the SPIR-V capability for descriptor heaps.

spvDeviceGroup

Represents the SPIR-V capability for DeviceGroup.

spvFloat8EXT

Represents the SPIR-V capability for using 8-bit floating point types.

spvFragmentBarycentricKHR

Represents the SPIR-V capability for using SPV_KHR_fragment_shader_barycentric.

spvFragmentFullyCoveredEXT

Represents the SPIR-V capability for using SPV_EXT_fragment_fully_covered functionality.

spvFragmentShaderPixelInterlockEXT

Represents the SPIR-V capability for using SPV_EXT_fragment_shader_interlock.

spvGroupNonUniform

Represents the SPIR-V GroupNonUniform capability (basic subgroup operations).

spvGroupNonUniformArithmetic

Represents the SPIR-V capability for group non-uniform arithmetic operations.

spvGroupNonUniformBallot

Represents the SPIR-V capability for group non-uniform ballot operations.

spvGroupNonUniformClustered

Represents the SPIR-V capability for group non-uniform clustered operations.

spvGroupNonUniformPartitionedNV

Represents the SPIR-V capability for group non-uniform partitioned operations.

spvGroupNonUniformQuad

Represents the SPIR-V capability for group non-uniform quad operations.

spvGroupNonUniformRotateKHR

Represents the SPIR-V capability for group non-uniform rotate operations.

spvGroupNonUniformShuffle

Represents the SPIR-V capability for group non-uniform shuffle operations.

spvGroupNonUniformShuffleRelative

Represents the SPIR-V capability for group non-uniform shuffle relative operations.

spvGroupNonUniformVote

Represents the SPIR-V capability for group non-uniform vote operations.

spvImageFootprintNV

Represents the SPIR-V capability for image footprint.

spvImageGatherExtended

Represents the SPIR-V capability for extended image gather operations.

spvImageQuery

Represents the SPIR-V capability for image query operations.

spvInt64Atomics

Represents the SPIR-V capability for 64-bit integer atomics.

spvMaximalReconvergenceKHR

Represents the SPIR-V capability for maximal reconvergence.

spvMeshShadingEXT

Represents the SPIR-V capability for mesh shading.

spvMinLod

Represents the SPIR-V capability for using minimum LOD operations.

spvQuadControlKHR

Represents the SPIR-V capability for quad group control.

spvRayQueryKHR

Represents the SPIR-V capability for ray query.

spvRayQueryPositionFetchKHR

Represents the SPIR-V capability for ray query position fetch.

spvRayTracingClusterAccelerationStructureNV

Represents the SPIR-V capability for cluster acceleration structure.

spvRayTracingKHR

Represents the SPIR-V capability for ray tracing.

spvRayTracingLinearSweptSpheresGeometryNV

Represents the SPIR-V capability for linear swept spheres.

spvRayTracingMotionBlurNV

Represents the SPIR-V capability for ray tracing motion blur.

spvRayTracingPositionFetchKHR

Represents the SPIR-V capability for ray tracing position fetch.

spvReplicatedCompositesEXT

Represents the SPIR-V capability for replicated composites

spvShaderClockKHR

Represents the SPIR-V capability for shader clock.

spvShaderInvocationReorderEXT

Represents the SPIR-V capability for shader invocation reorder (cross-vendor standard).

spvShaderInvocationReorderMotionEXT

Represents the SPIR-V capability for shader invocation reorder EXT + motion blur (cross-vendor SER with NV motion blur). Used for motion blur variants of EXT SER functions.

spvShaderInvocationReorderMotionNV

Represents the SPIR-V capability for shader invocation reorder + motion blur (NVIDIA-specific). Used for motion blur variants of SER functions that require both extensions.

spvShaderInvocationReorderNV

Represents the SPIR-V capability for shader invocation reorder (NVIDIA-specific). Inherits from spvShaderInvocationReorderEXT so that NV implies EXT capability.

spvShaderNonUniform

Represents the SPIR-V capability for non-uniform resource indexing.

spvShaderNonUniformEXT

Represents the SPIR-V capability for non-uniform resource indexing.

spvSparseResidency

Represents the SPIR-V capability for sparse residency.

spvTensorAddressingNV

Represents the SPIR-V capability for tensor addressing

spvVulkanMemoryModelDeviceScopeKHR

Represents the SPIR-V capability for vulkan memory model.

spvVulkanMemoryModelKHR

Represents the SPIR-V capability for vulkan memory model.

Compound Capabilities¶

Capabilities to specify capabilities created by other capabilities (raytracing, meshshading…)

amplification_mesh

Collection of shader stages

any_cpp_target

All “cpp syntax” code-gen targets

any_gfx_target

All slang-gfx compatible code-gen targets

any_stage

Collection of all shader stages

any_target

All code-gen targets

any_textual_target

All non-asm code-gen targets

anyhit_closesthit

Collection of shader stages

anyhit_closesthit_intersection

Collection of shader stages

anyhit_closesthit_intersection_miss

Collection of shader stages

anyhit_intersection

Collection of shader stages (for OptiX ObjectRayOrigin/Direction which excludes closesthit)

appendstructuredbuffer

Capabilities required to use AppendStructuredBuffer

atomic64

Capabilities needed for int64/uint64 atomic operations

atomic_bfloat16

Atomic operations on BFloat16 types. Requires SM 9.0 (Hopper) or higher on CUDA.

atomic_glsl

(GLSL/SPIRV) Capabilities required to use GLSL-400 atomic operations

atomic_glsl_float1

(GLSL/SPIRV) Capabilities required to use GLSL-tier-1 float-atomic operations

atomic_glsl_float2

(GLSL/SPIRV) Capabilities required to use GLSL-tier-2 float-atomic operations

atomic_glsl_halfvec

(GLSL/SPIRV) Capabilities required to use NVAPI GLSL-fp16 float-atomic operations

atomic_glsl_hlsl_cuda9_int64

(All implemented targets) Capabilities required to use atomic operations (cuda_sm_9 tier atomics)

atomic_glsl_hlsl_cuda_metal

(All implemented targets) Capabilities required to use atomic operations

atomic_glsl_hlsl_nvapi_cuda5_int64

(All implemented targets) Capabilities required to use atomic operations of int64 (cuda_sm_5 tier atomics)

atomic_glsl_hlsl_nvapi_cuda6_int64

(All implemented targets) Capabilities required to use atomic operations of int64 (cuda_sm_6 tier atomics)

atomic_glsl_hlsl_nvapi_cuda9_int64

(All implemented targets) Capabilities required to use atomic operations of int64 (cuda_sm_9 tier atomics)

atomic_glsl_hlsl_nvapi_cuda_metal_float1

(All implemented targets) Capabilities required to use atomic operations of GLSL tier-1 float atomics

atomic_glsl_int64

(GLSL/SPIRV) Capabilities required to use int64/uint64 atomic operations

atomic_hlsl

(hlsl only) Capabilities required to use hlsl atomic operations

atomic_hlsl_nvapi

(hlsl only) Capabilities required to use hlsl NVAPI atomics

atomic_hlsl_sm_6_6

(hlsl only) Capabilities required to use hlsl sm_6_6 atomics

atomic_reduce

Atomic reduction operations using PTX red instruction. Requires SM 7.0 on CUDA. On non-CUDA targets, falls back to regular atomic operations with no additional requirement.

atomicfloat

Capabilities needed to use GLSL-tier-1 float-atomic operations

atomicfloat2

Capabilities needed to use GLSL-tier-2 float-atomic operations

breakpoint

Capabilities required to enable shader breakpoints

bufferreference

Capabilities needed to use GLSL buffer-reference’s

bufferreference_int64

Capabilities needed to use GLSL buffer-reference’s with int64

byteaddressbuffer

Capabilities required to use ByteAddressBuffer

byteaddressbuffer_rw

Capabilities required to use RWByteAddressBuffer

compute_fragment

Collection of shader stages

compute_fragment_geometry_vertex

Collection of shader stages

compute_tesscontrol_tesseval

Collection of shader stages

consumestructuredbuffer

Capabilities required to use ConsumeStructuredBuffer

cooperative_matrix

Capabilities needed to use cooperative matrices

cooperative_matrix_2

Capabilities needed to use tensor addressing

cooperative_matrix_block_load

Capabilities needed to use decodeFunc with cooperative matrix load

cooperative_matrix_conversion

Capabilities needed to convert cooperative matrices, all the conversions can be supported by cuda

cooperative_matrix_map_element

Capabilities needed to use MapElement operation with cooperative matrix

cooperative_matrix_reduction

Capabilities needed to use reduction operations with cooperative matrix

cooperative_matrix_spirv

Capabilities needed to use reduction operations with cooperative matrix

cooperative_matrix_tensor_addressing

Capabilities needed to load or store with tensor_addressing extension

cooperative_vector

Capabilities needed to use cooperative vectors Note that cpp and cuda are supported via a fallback non-cooperative implementation No HLSL shader model bound yet

cooperative_vector_training

Capabilities needed to train cooperative vectors

cpp_cuda

CPP and CUDA code-gen targets

cpp_cuda_glsl_hlsl

CPP, CUDA, GLSL, and HLSL code-gen targets

cpp_cuda_glsl_hlsl_llvm

CPP, CUDA, GLSL, HLSL, and LLVM code-gen targets

cpp_cuda_glsl_hlsl_metal_spirv

CPP, CUDA, GLSL, HLSL, Metal and SPIRV code-gen targets

cpp_cuda_glsl_hlsl_metal_spirv_llvm

CPP, CUDA, GLSL, HLSL, Metal, SPIRV, and LLVM code-gen targets

cpp_cuda_glsl_hlsl_metal_spirv_wgsl

CPP, CUDA, GLSL, HLSL, Metal, SPIRV and WGSL code-gen targets

cpp_cuda_glsl_hlsl_metal_spirv_wgsl_llvm

CPP, CUDA, GLSL, HLSL, Metal, SPIRV, WGSL and LLVM code-gen targets

cpp_cuda_glsl_hlsl_spirv

CPP, CUDA, GLSL, HLSL, and SPIRV code-gen targets

cpp_cuda_glsl_hlsl_spirv_llvm

CPP, CUDA, GLSL, HLSL, SPIRV, and LLVM code-gen targets

cpp_cuda_glsl_hlsl_spirv_wgsl

CPP, CUDA, GLSL, HLSL, SPIRV and WGSL code-gen targets

cpp_cuda_glsl_hlsl_spirv_wgsl_llvm

CPP, CUDA, GLSL, HLSL, SPIRV, WGSL and LLVM code-gen targets

cpp_cuda_glsl_spirv

CPP, CUDA, GLSL and SPIRV code-gen targets

cpp_cuda_hlsl

CPP, CUDA, and HLSL code-gen targets

cpp_cuda_hlsl_metal_spirv

CPP, CUDA, HLSL, Metal, and SPIRV code-gen targets

cpp_cuda_hlsl_spirv

CPP, CUDA, HLSL, and SPIRV code-gen targets

cpp_cuda_llvm

CPP, CUDA and LLVM code-gen targets

cpp_cuda_metal_spirv

CPP, CUDA, Metal, and SPIRV code-gen targets

cpp_cuda_metal_spirv_llvm

CPP, CUDA, Metal, SPIRV, and LLVM code-gen targets

cpp_cuda_spirv

CPP, CUDA and SPIRV code-gen targets

cpp_cuda_spirv_llvm

CPP, CUDA, SPIRV and LLVM code-gen targets

cpp_glsl

CPP, and GLSL code-gen targets

cpp_glsl_hlsl_metal_spirv

CPP, GLSL, HLSL, Metal, and SPIRV code-gen targets

cpp_glsl_hlsl_metal_spirv_wgsl

CPP, GLSL, HLSL, Metal, SPIRV and WGSL code-gen targets

cpp_glsl_hlsl_spirv

CPP, GLSL, HLSL, and SPIRV code-gen targets

cpp_glsl_hlsl_spirv_wgsl

CPP, GLSL, HLSL, SPIRV and WGSL code-gen targets

cpp_hlsl

CPP, and HLSL code-gen targets

cpp_llvm

CPP and LLVM code-gen targets

cuda_glsl_hlsl

CUDA, GLSL, and HLSL code-gen targets

cuda_glsl_hlsl_metal_spirv

CUDA, GLSL, HLSL, Metal, and SPIRV code-gen targets

cuda_glsl_hlsl_metal_spirv_wgsl

CUDA, GLSL, HLSL, Metal, SPIRV and WGSL code-gen targets

cuda_glsl_hlsl_spirv

CUDA, GLSL, HLSL, and SPIRV code-gen targets

cuda_glsl_hlsl_spirv_llvm

CUDA, GLSL, HLSL, SPIRV, and LLVM code-gen targets

cuda_glsl_hlsl_spirv_wgsl

CUDA, GLSL, HLSL, SPIRV, and WGSL code-gen targets

cuda_glsl_metal_spirv

CUDA, GLSL, Metal, and SPIRV code-gen targets

cuda_glsl_metal_spirv_wgsl

CUDA, GLSL, Metal, SPIRV and WGSL code-gen targets

cuda_glsl_metal_spirv_wgsl_llvm

CUDA, GLSL, Metal, SPIRV, WGSL and LLVM code-gen targets

cuda_glsl_spirv

CUDA, GLSL, and SPIRV code-gen targets

cuda_hlsl

CUDA, and HLSL code-gen targets

cuda_hlsl_metal_spirv

CUDA, HLSL, Metal, and SPIRV code-gen targets

cuda_hlsl_spirv

CUDA, HLSL, SPIRV code-gen targets

cuda_metal_spirv

CUDA, Metal and SPIRV code-gen targets

cuda_spirv

CUDA and SPIRV code-gen targets

descriptor_handle

Targets that support DescriptorHandle types for bindless descriptor access.

domain_hull

Collection of shader stages

fragmentprocessing

Capabilities required to use fragment derivative operations (without GLSL derivativecontrol)

fragmentprocessing_derivativecontrol

Capabilities required to use fragment derivative operations (with GLSL derivativecontrol)

fragmentshaderbarycentric

Capabilities needed to use fragment-shader-barycentric’s

fragmentshaderinterlock

Capabilities needed for interlocked-fragment operations

getattributeatvertex

Capabilities required to use ‘getAttributeAtVertex’

glsl_hlsl_metal_spirv

GLSL, HLSL, Metal, and SPIRV code-gen targets

glsl_hlsl_metal_spirv_wgsl

GLSL, HLSL, Metal, SPIRV and WGSL code-gen targets

glsl_hlsl_spirv

GLSL, HLSL, and SPIRV code-gen targets

glsl_hlsl_spirv_wgsl

GLSL, HLSL, SPIRV and WGSL code-gen targets

glsl_metal_spirv

GLSL, Metal, and SPIRV code-gen targets

glsl_metal_spirv_wgsl

GLSL, Metal, SPIRV and WGSL code-gen targets

glsl_spirv

GLSL, and SPIRV code-gen targets

glsl_spirv_wgsl

GLSL, SPIRV, and WGSL code-gen targets

helper_lane

Capabilities required to enable helper-lane demotion

hlsl_spirv

HLSL, and SPIRV code-gen targets

image_loadstore

(GLSL/SPIRV) Capabilities required to use image load/image store operations

image_samples

Capabilities required to query image (RWTexture) sample info

image_size

Capabilities required to query image (RWTexture) size info

implicit_derivatives_sampling

Capabilities required for implicit derivatives sampling.

This capability is required by texture sampling functions such as _Texture.Sample() where the level of detail is determined by implicit derivatives.

@remark In GLSL, implicit level-of-detail texture() functions use the base texture when the implicit derivatives are unavailable. When this capability is not present, invocations of these functions are translated to invocations of _Texture.SampleLevelZero().

@remark Implicit derivatives for the compute stage can be enabled by declaring capability GL_NV_compute_shader_derivatives (GLSL), SPV_KHR_compute_shader_derivatives (SPIR-V), or profile cs_6_6 (HLSL).

mem_model

Capabilities needed to use memory model

memorybarrier

Capabilities required to use sm_5_0 style memory barriers

meshshading

Ccapabilities required to use mesh shading features

motionblur_nv

Capabilities needed for raytracing-motionblur (NVIDIA-specific)

nonuniformqualifier

Capabilities required to use NonUniform qualifier

nvapi

NVAPI capability for HLSL

pack_vector

Capabilities required to use pack/unpack intrinsics on packed vector data

printf

Capabilities required to use ‘printf’

quad_control

Capabilities required to enable quad group control

raygen_closesthit_miss

Collection of shader stages

raygen_closesthit_miss_callable

Collection of shader stages

rayquery

Capabilities needed for compute-shader rayquery

rayquery_position

Collection of capabilities for rayquery + ray_tracing_position_fetch.

raytracing

Capabilities needed for minimal raytracing support

raytracing_allstages

Collection of capabilities for raytracing with all raytracing stages.

raytracing_anyhit

Collection of capabilities for raytracing with the shader stage of anyhit.

raytracing_anyhit_closesthit

Collection of capabilities for raytracing with the shader stages of anyhit and closesthit.

raytracing_anyhit_closesthit_intersection

Collection of capabilities for raytracing with the shader stages of anyhit, closesthit, and intersection.

raytracing_anyhit_closesthit_intersection_miss

Collection of capabilities for raytracing with the shader stages of anyhit, closesthit, intersection, and miss.

raytracing_intersection

Collection of capabilities for raytracing with the shader stage of intersection.

raytracing_lss

Collection of capabilities for linear swept spheres.

raytracing_lss_ho

hit object APIs allow raygen shaders, but not the non-hit object APIs. So we have this special capdef specifically for the hitobject variant. Collection of capabilities for linear swept spheres.

raytracing_motionblur

Capabilities needed for compute-shader rayquery and motion-blur

raytracing_motionblur_anyhit_closesthit_intersection_miss

Collection of capabilities for raytracing + motion blur and the shader stages of anyhit, closesthit, intersection, and miss.

raytracing_motionblur_raygen_closesthit_miss

Collection of capabilities for raytracing + motion blur and the shader stages of raygen, closesthit, and miss.

raytracing_object_space_ray

Collection of capabilities for ObjectRayOrigin/ObjectRayDirection. CUDA/OptiX only supports anyhit and intersection stages, while other targets also support closesthit.

raytracing_position

Collection of capabilities for raytracing + ray_tracing_position_fetch and the shader stages of anyhit and closesthit.

raytracing_raygen_closesthit_miss

Collection of capabilities for raytracing with the shader stages of raygen, closesthit, and miss.

raytracing_raygen_closesthit_miss_callable

Collection of capabilities for raytracing the shader stages of raygen, closesthit, miss, and callable.

raytracing_stages

Collection of shader stages

raytracingstages_compute

Collection of shader stages

raytracingstages_compute_amplification_mesh

Collection of shader stages

raytracingstages_compute_fragment

Collection of shader stages

raytracingstages_compute_fragment_geometry_vertex

Collection of shader stages

raytracingstages_fragment

Collection of shader stages

ser

Capabilities needed for shader-execution-reordering (all paths) Includes NVIDIA-specific (NV), cross-vendor standard (EXT), DXR 1.3 native, and CUDA paths

ser_any_closesthit_intersection_miss

Collection of capabilities for raytracing + shader execution reordering and the shader stages of anyhit, closesthit, intersection, and miss.

ser_anyhit_closesthit

Collection of capabilities for raytracing + shader execution reordering and the shader stages of anyhit and closesthit.

ser_anyhit_closesthit_intersection

Collection of capabilities for raytracing + shader execution reordering and the shader stages of anyhit, closesthit, and intersection.

ser_dxr

Capabilities needed for shader-execution-reordering (native DXR 1.3 path)

ser_dxr_raygen

Collection of capabilities for DXR 1.3 native SER (HLSL only) with raygen stage.

ser_dxr_raygen_closesthit_miss

Collection of capabilities for DXR 1.3 native SER (HLSL only) with raygen, closesthit, miss stages.

ser_motion

Capabilities needed for shader-execution-reordering and motion-blur

ser_motion_raygen

Collection of capabilities for raytracing raytracing + motion blur + shader execution reordering and the shader stage of raygen.

ser_motion_raygen_closesthit_miss

Collection of capabilities for raytracing + motion blur + shader execution reordering and the shader stages of raygen, closesthit, and miss.

ser_nv

NVIDIA-specific SER capabilities (excludes cross-vendor EXT) Includes GL_NV_shader_invocation_reorder (GLSL/SPIRV) and CUDA paths

ser_nv_motion

NVIDIA-specific SER + motion blur

ser_nv_motion_raygen_closesthit_miss

NVIDIA-specific SER + motion blur for raygen, closesthit, miss stages

ser_nv_raygen

NVIDIA-specific SER for raygen stage (GLSL/SPIRV NV paths)

ser_nv_raygen_closesthit_miss

NVIDIA-specific SER for raygen, closesthit, miss stages (GLSL/SPIRV NV paths)

ser_nvapi

Capabilities needed for shader-execution-reordering (NVAPI path for HLSL)

ser_raygen

Collection of capabilities for raytracing + shader execution reordering and the shader stage of raygen.

ser_raygen_closesthit_miss

Collection of capabilities for raytracing + shader execution reordering and the shader stages of raygen, closesthit, and miss.

shader5_sm_4_0

Capabilities required to use sm_4_0 features apart of GL_ARB_gpu_shader5

shader5_sm_5_0

Capabilities required to use sm_5_0 features apart of GL_ARB_gpu_shader5

shaderclock

Capabilities needed for realtime clocks

shaderinvocationgroup

Capabilities required to use GLSL-style subgroup operations ‘subgroup_vote’

shadermemorycontrol

(gfx targets) Capabilities needed to use memory barriers

shadermemorycontrol_compute

(gfx targets) Capabilities required to use memory barriers that only work for raytracing & compute shader stages

structuredbuffer

Capabilities required to use StructuredBuffer

structuredbuffer_rw

Capabilities required to use RWStructuredBuffer

subgroup_arithmetic

Capabilities required to use GLSL-style subgroup operations ‘subgroup_arithmetic’

subgroup_ballot

Capabilities required to use GLSL-style subgroup operations ‘subgroup_ballot’

subgroup_ballot_activemask

Capabilities required to use GLSL-style subgroup operations ‘subgroup_ballot_activemask’

subgroup_basic

Capabilities required to use GLSL-style subgroup operations ‘subgroup_basic’

subgroup_basic_ballot

Capabilities required to use GLSL-style subgroup operations ‘subgroup_basic_ballot’

subgroup_clustered

Capabilities required to use GLSL-style subgroup operations ‘subgroup_clustered’

subgroup_partitioned

Capabilities required to use GLSL-style subgroup operations ‘subgroup_partitioned’

subgroup_quad

Capabilities required to use GLSL-style subgroup operations ‘subgroup_quad’

subgroup_rotate

Capabilities required to use GLSL-style subgroup rotate operations ‘subgroup_rotate’

subgroup_shuffle

Capabilities required to use GLSL-style subgroup operations ‘subgroup_shuffle’

subgroup_shufflerelative

Capabilities required to use GLSL-style subgroup operations ‘subgroup_shuffle_relative’

subgroup_vote

Capabilities required to use GLSL-style subgroup operations ‘subgroup_vote’

subpass

Capabilities required to use Subpass-Input’s

tensor_addressing

Capabilities needed to use tensor addressing

texture_gather

Capabilities required to use ‘vertex/fragment/geometry shader only’ texture gather operations

texture_querylevels

Capabilities required to query texture level info

texture_querylod

Capabilities required to query texture LOD info

texture_shadowgrad

Capabilities required for shadow texture sampling with bias and gradients. New in HLSL SM6.8 but existed in older GLSL and SPIRV targets.

texture_shadowlod

Capabilities required to query shadow texture lod info

texture_size

Capabilities required to query texture sample info

texture_sm_4_0

Capabilities required to use sm_4_0 texture operations

texture_sm_4_0_fragment

Capabilities required to use ‘fragment shader only’ texture operations

texture_sm_4_1

Capabilities required to use sm_4_1 texture operations

texture_sm_4_1_clamp_fragment

Capabilities required to use ‘fragment shader only’ texture clamp operations

texture_sm_4_1_compute_fragment

Capabilities required to use ‘compute/fragment shader only’ texture operations. We do not require ‘compute’/‘fragment’ shader capabilities since this seems to be incorrect behavior despite what official documentation says.

texture_sm_4_1_samplerless

Capabilities required to use sm_4_1 samplerless texture operations

texture_sm_4_1_vertex_fragment_geometry

Capabilities required to use ‘fragment/geometry shader only’ texture clamp operations

texturefootprint

Capabilities required to use basic TextureFootprint operations

texturefootprintclamp

Capabilities required to use TextureFootprint clamp operations

vk_mem_model

Capabilities needed to use vulkan memory model

Other¶

Capabilities which may be deprecated

DX_4_0

Use sm_4_0 instead

DX_4_1

Use sm_4_1 instead

DX_5_0

Use sm_5_0 instead

DX_5_1

Use sm_5_1 instead

DX_6_0

Use sm_6_0 instead

DX_6_1

Use sm_6_1 instead

DX_6_10

Use sm_6_10 instead

DX_6_2

Use sm_6_2 instead

DX_6_3

Use sm_6_3 instead

DX_6_4

Use sm_6_4 instead

DX_6_5

Use sm_6_5 instead

DX_6_6

Use sm_6_6 instead

DX_6_7

Use sm_6_7 instead

DX_6_8

Use sm_6_8 instead

DX_6_9

Use sm_6_9 instead

GLSL_410_SPIRV_1_0

User should not use this capability

GLSL_420_SPIRV_1_0

User should not use this capability

GLSL_430_SPIRV_1_0

User should not use this capability

GLSL_430_SPIRV_1_0_compute

User should not use this capability

METAL_2_3

Use metallib_2_3 instead

METAL_2_4

Use metallib_2_4 instead

METAL_3_0

Use metallib_3_0 instead

METAL_3_1

Use metallib_3_1 instead

METAL_4_0

Use metallib_4_0 instead

SPIRV_1_0

Use spirv_1_0 instead

SPIRV_1_1

Use spirv_1_1 instead

SPIRV_1_2

Use spirv_1_2 instead

SPIRV_1_3

Use spirv_1_3 instead

SPIRV_1_4

Use spirv_1_4 instead

SPIRV_1_5

Use spirv_1_5 instead

SPIRV_1_6

Use spirv_1_6 instead

SPV_KHR_variable_pointers

Represents the SPIRV-V Variable Pointers extension.

all

User should not use this capability

optix_coopvec

Represents capabilities required for optix cooperative vector support.

optix_multilevel_traversal

Represents capabilities required for optix multi-level traversal support.