VulkanFeatures.md

Source folder: Vulkan backend

Supported extensions:

• VK_EXT_swapchain_colorspace
• VK_EXT_validation_features
• VK_KHR_descriptor_update_template
• VK_KHR_sampler_ycbcr_conversion
• VK_KHR_draw_indirect_count
• VK_EXT_descriptor_indexing
• VK_KHR_timeline_semaphore
• VK_KHR_buffer_device_address
• VK_EXT_sampler_filter_minmax
• VK_EXT_host_query_reset
• VK_KHR_image_format_list
• VK_KHR_synchronization2
• VK_EXT_load_store_op_none
• VK_EXT_mesh_shader
• VK_KHR_global_priority
• VK_EXT_calibrated_timestamps
• VK_KHR_acceleration_structure
• VK_KHR_ray_tracing_pipeline
• VK_KHR_ray_query
• VK_KHR_ray_tracing_maintenance1
• VK_EXT_vertex_attribute_divisor, VK_KHR_vertex_attribute_divisor
• VK_EXT_memory_budget
• VK_KHR_fragment_shading_rate
• VK_KHR_get_surface_capabilities2
• VK_EXT_swapchain_colorspace
• VK_EXT_debug_utils
• VK_EXT_validation_features
• VK_KHR_dedicated_allocation
• VK_KHR_create_renderpass2
• VK_KHR_sampler_mirror_clamp_to_edge
• VK_EXT_texture_compression_astc_hdr
• VK_EXT_texel_buffer_alignment
• VK_KHR_portability_subset
• VK_AMD_shader_core_properties, VK_AMD_shader_core_properties2
• VK_KHR_pipeline_executable_properties
• VK_EXT_device_memory_report
• VK_NV_ray_tracing_validation
• VK_NV_clip_space_w_scaling
• VK_EXT_subgroup_size_control

In progress:

• VK_KHR_video_queue
• VK_KHR_video_encode_queue
• VK_KHR_video_decode_queue
• VK_KHR_video_decode_h264
• VK_KHR_video_decode_h265
• VK_EXT_video_encode_h264
• VK_EXT_video_encode_h265
• VK_KHR_external_memory, VK_KHR_external_memory_capabilities
• VK_KHR_external_semaphore, VK_KHR_external_semaphore_fd, VK_KHR_external_semaphore_capabilities
• VK_EXT_queue_family_foreign
• VK_ANDROID_external_memory_android_hardware_buffer (for NdkCamera)

Pending extensions with high priority:

• VK_EXT_pageable_device_local_memory

Pending extensions:

• VK_KHR_multiview
• VK_KHR_uniform_buffer_standard_layout
• VK_KHR_format_feature_flags2
• VK_EXT_inline_uniform_block
• VK_EXT_depth_range_unrestricted
• VK_EXT_sample_locations
• VK_EXT_rasterization_order_attachment_access
• VK_EXT_astc_decode_mode
• VK_EXT_image_compression_control, VK_EXT_image_compression_control_swapchain
• VK_NV_shader_image_footprint
• VK_NV_device_generated_commands, VK_NV_device_generated_commands_compute
• VK_ARM_rasterization_order_attachment_access
• VK_KHR_incremental_present
• VK_KHR_present_id
• VK_KHR_present_wait
• VK_EXT_full_screen_exclusive
• VK_EXT_host_image_copy
• VK_KHR_maintenance5
• VK_NV_memory_decompression
• VK_EXT_memory_priority
• VK_NV_external_memory_rdma
• VK_EXT_direct_mode_display
• VK_EXT_hdr_metadata
• VK_EXT_headless_surface ???
• VK_EXT_metal_objects
• VK_EXT_multisampled_render_to_single_sampled
• VK_EXT_attachment_feedback_loop_layout
• VK_EXT_depth_clamp_zero_one
• VK_EXT_subpass_merge_feedback
• VK_NV_representative_fragment_test
• VK_QCOM_tile_properties
• VK_EXT_display_control
• VK_EXT_display_surface_counter
• VK_EXT_opacity_micromap, GL_EXT_opacity_micromap
• VK_EXT_pipeline_protected_access
• VK_NV_present_barrier
• VK_NV_shader_subgroup_partitioned
• VK_EXT_descriptor_buffer
• VK_EXT_swapchain_maintenance1
• VK_EXT_surface_maintenance1
• VK_ARM_scheduling_controls
• VK_KHR_ray_tracing_position_fetch, GL_EXT_ray_tracing_position_fetch
• VK_EXT_discard_rectangles
• VK_EXT_image_sliced_view_of_3d
• VK_EXT_image_2d_view_of_3d
• VK_EXT_image_view_min_lod
• VK_EXT_line_rasterization
• VK_EXT_nested_command_buffer
• VK_EXT_post_depth_coverage
• VK_ARM_render_pass_striped
• VK_NV_dedicated_allocation_image_aliasing
• VK_NV_extended_sparse_address_space
• VK_ARM_shader_core_properties
• VK_KHR_dynamic_rendering_local_read
• VK_KHR_line_rasterization
• VK_KHR_maintenance6
• VK_KHR_shader_expect_assume
• VK_KHR_shader_float_controls2
• VK_KHR_shader_subgroup_rotate
• VK_KHR_video_maintenance1
• VK_NV_extended_sparse_address_space
• VK_EXT_shader_tile_image
• VK_EXT_primitive_topology_list_restart
• VK_NV_viewport_array2, GL_NV_viewport_array2 - add gl_ViewportMask
• VK_NV_ray_tracing_invocation_reorder, GL_NV_shader_invocation_reorder
• VK_NV_displacement_micromap, GL_NV_displacement_micromap
• VK_KHR_push_descriptor
• VK_KHR_shader_integer_dot_product - ML
• VK_KHR_pipeline_binary

Pending for debugging:

• VK_EXT_device_fault
• VK_EXT_robustness2
• VK_EXT_device_address_binding_report
• VK_EXT_frame_boundary
• VK_EXT_layer_settings
• VK_EXT_pipeline_robustness
• VK_NV_device_diagnostics_config
• VK_NV_device_diagnostic_checkpoints

Pending for profiling:

• VK_EXT_pipeline_creation_feedback
• VK_EXT_primitives_generated_query
• VK_KHR_performance_query

Will not be supported

• VK_NV_linear_color_attachment
• VK_EXT_non_seamless_cube_map
• VK_NV_geometry_shader_passthrough
• VK_NV_compute_shader_derivatives, GL_NV_compute_shader_derivatives
• VK_KHR_pipeline_library
• VK_NV_raw_access_chains - for HLSL
• VK_KHR_index_type_uint8
• VK_NV_copy_memory_indirect

Supported shader extensions:

• VK_KHR_shader_clock, GL_EXT_shader_realtime_clock, GL_ARB_shader_clock
• GL_EXT_control_flow_attributes
• GL_EXT_debug_printf
• GL_EXT_samplerless_texture_functions
• GL_KHR_shader_subgroup_*
• VK_KHR_shader_subgroup_extended_types, GL_EXT_shader_subgroup_extended_types_*
• VK_KHR_8bit_storage, GL_EXT_shader_8bit_storage
• VK_KHR_shader_float16_int8, GL_EXT_shader_explicit_arithmetic_types_*
• VK_KHR_16bit_storage, GL_EXT_shader_16bit_storage
• GL_ARB_gpu_shader_int64
• GL_ARB_gpu_shader_fp64
• GL_EXT_buffer_reference, GL_EXT_buffer_reference2, GL_EXT_buffer_reference_uvec2
• VK_KHR_shader_atomic_int64, GL_EXT_shader_atomic_int64
• GL_EXT_shader_image_int64
• VK_EXT_shader_atomic_float, GL_EXT_shader_atomic_float
• VK_EXT_shader_atomic_float2, GL_EXT_shader_atomic_float2
• GL_EXT_mesh_shader
• GL_EXT_ray_tracing, GL_EXT_ray_flags_primitive_culling, GL_EXT_ray_cull_mask
• GL_EXT_ray_query
• VK_EXT_shader_viewport_index_layer, GL_ARB_shader_viewport_layer_array
• GL_ARB_sparse_texture_clamp
• VK_NV_shader_sm_builtins, GL_NV_shader_sm_builtins
• VK_ARM_shader_core_builtins, GL_ARM_shader_core_builtins
• GL_EXT_nonuniform_qualifier
• VK_KHR_vulkan_memory_model, GL_KHR_memory_scope_semantics
• VK_EXT_shader_demote_to_helper_invocation, GL_EXT_demote_to_helper_invocation
• GL_EXT_fragment_shading_rate
• VK_EXT_fragment_shader_interlock, GL_ARB_fragment_shader_interlock
• VK_KHR_shader_draw_parameters, GL_ARB_shader_draw_parameters
• VK_KHR_fragment_shader_barycentric, GL_EXT_fragment_shader_barycentric
• VK_KHR_cooperative_matrix, GL_KHR_cooperative_matrix (deprecated: VK_NV_cooperative_matrix, GL_NV_cooperative_matrix, GL_NV_integer_cooperative_matrix)
• VK_KHR_shader_subgroup_uniform_control_flow, GL_EXT_subgroup_uniform_control_flow
• GL_EXT_control_flow_attributes2
• VK_KHR_shader_quad_control, GL_EXT_shader_quad
• VK_KHR_shader_maximal_reconvergence, GL_EXT_maximal_reconvergence

Pending shader extensions:

• GL_EXT_multiview
• GL_EXT_null_initializer
• VK_KHR_zero_initialize_workgroup_memory
• VK_EXT_scalar_block_layout, GL_EXT_scalar_block_layout
• VK_HUAWEI_subpass_shading, GL_HUAWEI_subpass_shading ???
• VK_KHR_shader_terminate_invocation
• VK_KHR_workgroup_memory_explicit_layout
• VK_EXT_shader_image_atomic_int64
• VK_HUAWEI_cluster_culling_shader, GL_HUAWEI_cluster_culling_shader
• GL_EXT_expect_assume - depends on compiler?
• VK_NV_shader_atomic_float16_vector
• VK_KHR_shader_float_controls, SPV_KHR_float_controls - not supported by glslang
• VK_KHR_shader_float_controls2, SPV_KHR_float_controls2 - not supported by glslang
• GL_EXT_spirv_intrinsics
• GL_EXT_shared_memory_block
• GL_EXT_fragment_invocation_density
• GL_EXT_shader_tile_image
• GL_EXT_subgroupuniform_qualifier - not supported by glslang
• GL_EXT_terminate_invocation
• GL_NV_ray_tracing_motion_blur
• GL_NV_shader_texture_footprint
• GL_NV_shader_subgroup_partitioned

Will not be supported

• GL_EXT_vulkan_glsl_relaxed
• GL_NV_fragment_shader_barycentric
• GL_NV_mesh_shader
• GL_NV_primitive_shading_rate
• GL_NV_ray_tracing
• GL_NV_shading_rate_image
• GL_NVX_multiview_per_view_attributes
• GL_QCOM_image_processing

Features:

• Async compute and transfer queues
• Async queues with concurrent sharing mode
• Async queues with exclusive sharing mode
• Push constants (compatible with Metal)
• Custom allocator for GPU memory (fast linear allocator, block allocator)
• Shader subgroup
• Secondary command buffers for async recording of draw commands
• Immutable samplers

TODO:

• sparse memory
• reuse device memory immediately, destroy object with delay - memory on GPU side will be synchronized
• unbounded array descriptors
• push descriptors