Declaration of AcceleratorBuffer. More...

#include "AlignedAllocator.h"
#include "QueueManager.h"
#include <sys/mman.h>
#include <unistd.h>
#include <type_traits>

Include dependency graph for AcceleratorBuffer.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	AcceleratorBufferIterator< T >

class	AcceleratorBuffer< T >
	A buffer of memory which may be used by GPU acceleration. More...

Typedefs
typedef bool(*	MemoryPressureHandler) (MemoryPressureLevel level, MemoryPressureType type, size_t requestedSize)
	Memory pressure handler type, called when free memory reaches a warning level or a Vulkan allocation fails.

Enumerations
enum class	MemoryPressureLevel { Hard , Soft }
	Levels of memory pressure. More...

enum class	MemoryPressureType { Host , Device }
	Types of memory pressure. More...

Functions
bool	OnMemoryPressure (MemoryPressureLevel level, MemoryPressureType type, size_t requestedSize)
	Called when we run low on memory. More...

template<class T >
std::ptrdiff_t	operator- (const AcceleratorBufferIterator< T > &a, const AcceleratorBufferIterator< T > &b)

Variables
uint32_t	g_vkPinnedMemoryType
	Vulkan memory type for CPU-based memory that is also GPU-readable.

uint32_t	g_vkLocalMemoryType
	Vulkan memory type for GPU-based memory (generally not CPU-readable, except on unified memory systems)

std::shared_ptr< vk::raii::Device >	g_vkComputeDevice
	The Vulkan device selected for compute operations (may or may not be same device as rendering)

std::unique_ptr< vk::raii::CommandBuffer >	g_vkTransferCommandBuffer
	Command buffer for AcceleratorBuffer transfers. More...

std::shared_ptr< QueueHandle >	g_vkTransferQueue
	Queue for AcceleratorBuffer transfers. More...

std::mutex	g_vkTransferMutex
	Mutex for interlocking access to g_vkTransferCommandBuffer and g_vkTransferCommandPool.

bool	g_hasDebugUtils
	Indicates whether the VK_EXT_debug_utils extension is available.

bool	g_vulkanDeviceHasUnifiedMemory
	Indicates whether the Vulkan device is unified memory. More...

std::set< MemoryPressureHandler >	g_memoryPressureHandlers
	List of handlers for low memory registered by various subsystems.

Detailed Description

Declaration of AcceleratorBuffer.

Author: Andrew D. Zonenberg

Enumeration Type Documentation

◆ MemoryPressureLevel

enum class MemoryPressureLevel

strong

Levels of memory pressure.

Enumerator

Hard

A memory allocation has failed and we need to free memory immediately to continue execution.

Soft

Free memory has reached a warning threshold.

We should trim caches or otherwise try to make space but don't need to be too aggressive about it.

This level is only available if we have VK_EXT_memory_budget; without this extension we do not know about memory pressure until a hard allocation failure occurs.

◆ MemoryPressureType

enum class MemoryPressureType

strong

Types of memory pressure.

Enumerator
Host	Pinned CPU-side memory.
Device	GPU-side memory.

Function Documentation

◆ OnMemoryPressure()

bool OnMemoryPressure	(	MemoryPressureLevel	level,
		MemoryPressureType	type,
		size_t	requestedSize
	)

Called when we run low on memory.

Parameters

level	Indicates if this is a soft or hard memory exhaustion condition
type	Indicates if we are low on CPU or GPU memory
requestedSize	For hard memory exhaustion, the size of the failing allocation. For soft exhaustion, ignored and set to zero

Returns: True if memory was freed, false if no space could be freed

Classes

Typedefs

Enumerations