Inherits AZ::RHI::ResourcePoolDescriptor.
Inherited by AZ::DX12::BufferPoolDescriptor, and AZ::Metal::BufferPoolDescriptor.
Public Member Functions | |
| AZ_RTTI (BufferPoolDescriptor, "{53074556-41D1-4246-8BF0-E5B096997C65}", ResourcePoolDescriptor) | |
 Public Member Functions inherited from AZ::RHI::ResourcePoolDescriptor | |
| AZ_RTTI (ResourcePoolDescriptor, "{C4B9BF83-B171-4DB9-93D6-0879C7CEF5C2}") | |
| AZ_CLASS_ALLOCATOR (ResourcePoolDescriptor, SystemAllocator) | |
Static Public Member Functions | |
| static void | Reflect (AZ::ReflectContext *context) |
| Static Public Member Functions inherited from AZ::RHI::ResourcePoolDescriptor | |
| static void | Reflect (AZ::ReflectContext *context) |
Public Attributes | |
| HeapMemoryLevel | m_heapMemoryLevel = HeapMemoryLevel::Device |
| HostMemoryAccess | m_hostMemoryAccess = HostMemoryAccess::Write |
| BufferBindFlags | m_bindFlags = BufferBindFlags::None |
| AZ::u64 | m_largestPooledAllocationSizeInBytes = 0 |
| HardwareQueueClassMask | m_sharedQueueMask = HardwareQueueClassMask::All |
| The mask of queue classes supporting shared access of this pool. | |
| Public Attributes inherited from AZ::RHI::ResourcePoolDescriptor | |
| AZ::u64 | m_budgetInBytes = 0 |
Member Data Documentation
◆ m_bindFlags
| BufferBindFlags AZ::RHI::BufferPoolDescriptor::m_bindFlags = BufferBindFlags::None |
The set of buffer bind flags supported by this pool. Buffers must be initialized with the exact bind flags as the parent pool.
◆ m_heapMemoryLevel
| HeapMemoryLevel AZ::RHI::BufferPoolDescriptor::m_heapMemoryLevel = HeapMemoryLevel::Device |
The memory heap to store final buffer allocations. Currently, there are two supported options:
- 'Device' memory is stored local to GPU hardware. On certain platforms this may require a transfer from Host memory through a DMA controller. In this scenario, Host memory will be used to stage the transfer.
- 'Host' memory is stored local to the CPU. This guarantees fast CPU access and will not require an explicit staging transfer. However, GPU read access will be slower as a result.
◆ m_hostMemoryAccess
| HostMemoryAccess AZ::RHI::BufferPoolDescriptor::m_hostMemoryAccess = HostMemoryAccess::Write |
If the heap memory type is host memory, this specifies the CPU access pattern from the user in system memory.
- Host 'Write' pools are written by the CPU and read by the GPU.
- Host 'Read' pools are written by the GPU and read by the CPU.
If the Device heap is used, this value must be 'Write', as only 'Write' operations are permitted for Map operations. Attempting to assign 'Read' when the 'Device' heap is specified will result in an error.
◆ m_largestPooledAllocationSizeInBytes
| AZ::u64 AZ::RHI::BufferPoolDescriptor::m_largestPooledAllocationSizeInBytes = 0 |
Specifies the largest allocation required of the pool. Useful if you are allocating buffers larger than the RHI default page size.
The documentation for this class was generated from the following file:
- Gems/Atom/RHI/Code/Include/Atom/RHI.Reflect/BufferPoolDescriptor.h
