Public Attributes
bool	m_geometryShader
	Whether the adapter supports geometry shaders.

bool	m_computeShader
	Whether the adapter supports compute shaders.

bool	m_independentBlend
	Whether color attachments can utilize independent blend modes.

bool	m_dualSourceBlending = false
	Whether the adapter supports dual source blending.

bool	m_customSamplePositions = false
	Whether the adapter supports custom sample positions.

AZStd::array< QueryTypeFlags, HardwareQueueClassCount >	m_queryTypesMask = { { QueryTypeFlags::None, QueryTypeFlags::None, QueryTypeFlags::None } }
	Supported query types.

bool	m_occlusionQueryPrecise = false
	Whether the adapter supports precise occlusion query.

bool	m_predication = false
	Whether the adapter supports predication.

bool	m_indirectDrawSupport = true
	Whether the adapter supports indirect draw.

bool	m_indirectDrawCountBufferSupported = false
	Whether the adapter supports a count buffer when doing indirect drawing.

bool	m_indirectDispatchCountBufferSupported = false
	Whether the adapter supports a count buffer when doing indirect dispatching.

bool	m_indirectDrawStartInstanceLocationSupported = false
	Whether the adapter supports supplying a start instance location when doing indirect drawing.

RHI::IndirectCommandTiers	m_indirectCommandTier = RHI::IndirectCommandTiers::Tier0
	Tier of supported indirect commands.

SubpassInputSupportType	m_subpassInputSupport = SubpassInputSupportType::None
	The type of support for subpass inputs of render target attachments.

bool	m_rayTracing = false
	Whether Ray Tracing support is available.

bool	m_rayTracingClas = false
	Whether Ray Tracing supports CLAS structures.

bool	m_unboundedArrays = false
	Whether Unbounded Array support is available.

bool	m_simulateBindlessUA = false
	Whether Unbounded Array support is simulated for Bindless Srg.

bool	m_isPsoCacheFileOperationsNeeded = true
	Whether PipelineLibrary related serialized data needs to be loaded/saved explicitly as drivers (like dx12/vk) do not support it internally.

bool	m_tiledResource = true
	Whether supports tile resource.

RHI::ShadingRateTypeFlags	m_shadingRateTypeMask = RHI::ShadingRateTypeFlags::None
	Supported modes for setting the variable shading rate.

RHI::ShadingRateFlags	m_shadingRateMask = RHI::ShadingRateFlags::None
	Supported values for setting the shading rate.

bool	m_dynamicShadingRateImage = false

bool	m_waveOperation = false
	Whether the adapter supports wave/subgroup operation.

RHI::ScalingFlags	m_swapchainScalingFlags = RHI::ScalingFlags::None
	Whether swapchain scaling support is available.

bool	m_signalFenceFromCPU = false

bool	m_float16 = false
	Whether float16 (half-precision floating-point format) support is available.

bool	m_crossDeviceDeviceMemory = false
	Whether device memory can be shared with other devices.

bool	m_crossDeviceHostMemory = false
	Whether host memory can be shared with other devices.

bool	m_crossDeviceFences = false
	Whether fences can be shared with other devices.

Member Data Documentation

◆ m_dynamicShadingRateImage

bool AZ::RHI::DeviceFeatures::m_dynamicShadingRateImage = false

Whether the shading rate attachment is read by the GPU during execution. Otherwise, the image is read by the CPU during scope execution (so it must not be modified by either CPU or GPU at that time).

◆ m_signalFenceFromCPU

bool AZ::RHI::DeviceFeatures::m_signalFenceFromCPU = false

Whether fences can be signalled from the CPU If this is false, the Fence::SignalOnCpu inserts the signal command into a queue

The documentation for this struct was generated from the following file:

Gems/Atom/RHI/Code/Include/Atom/RHI.Reflect/DeviceFeatures.h

Public Attributes

Member Data Documentation

◆ m_dynamicShadingRateImage

◆ m_signalFenceFromCPU