Compute-specific pipeline and dispatch orchestration. More...

#include <ComputePress.hpp>

Collaboration diagram for MayaFlux::Portal::Graphics::ComputePress:

Classes
struct	PipelineState

Public Member Functions
	ComputePress (const ComputePress &)=delete

ComputePress &	operator= (const ComputePress &)=delete

	ComputePress (ComputePress &&) noexcept=delete

ComputePress &	operator= (ComputePress &&) noexcept=delete

bool	initialize ()

void	shutdown ()

ComputePipelineID	create_pipeline (ShaderID shader_id, const std::vector< std::vector< DescriptorBindingConfig > > &descriptor_sets={}, size_t push_constant_size=0)
	Create compute pipeline.

ComputePipelineID	create_pipeline_auto (ShaderID shader_id, size_t push_constant_size=0)
	Create pipeline with auto-reflection.

void	destroy_pipeline (ComputePipelineID pipeline_id)

void	bind_pipeline (CommandBufferID cmd_id, ComputePipelineID pipeline_id)
	Bind pipeline to active command buffer.

void	bind_descriptor_sets (CommandBufferID cmd_id, ComputePipelineID pipeline_id, const std::vector< DescriptorSetID > &descriptor_set_ids)
	Bind descriptor sets to active command buffer.

void	push_constants (CommandBufferID cmd_id, ComputePipelineID pipeline_id, const void *data, size_t size)
	Push constants to active command buffer.

void	dispatch (CommandBufferID cmd_id, uint32_t x, uint32_t y, uint32_t z)
	Dispatch compute workgroups.

void	dispatch_indirect (CommandBufferID cmd_id, vk::Buffer indirect_buffer, vk::DeviceSize offset=0)
	Dispatch compute workgroups indirectly.

std::vector< DescriptorSetID >	allocate_pipeline_descriptors (ComputePipelineID pipeline_id)
	All-in-one: allocate descriptors for pipeline.

void	bind_all (CommandBufferID cmd_id, ComputePipelineID pipeline_id, const std::vector< DescriptorSetID > &descriptor_set_ids, const void *push_constants_data=nullptr, size_t push_constant_size=0)
	All-in-one: bind pipeline + descriptors + push constants.

Static Public Member Functions
static ComputePress &	instance ()

Private Member Functions
	ComputePress ()=default

	~ComputePress ()

Private Attributes
std::shared_ptr< Core::VKDescriptorManager >	m_descriptor_manager

std::unordered_map< ComputePipelineID, PipelineState >	m_pipelines

std::atomic< uint64_t >	m_next_pipeline_id { 1 }

ShaderFoundry *	m_shader_foundry = nullptr

Static Private Attributes
static bool	s_initialized = false

Detailed Description

Compute-specific pipeline and dispatch orchestration.

Responsibilities:

Create compute pipelines
Bind compute pipelines
Dispatch compute workgroups (direct + indirect)
Compute-specific optimizations

Uses ShaderFoundry for:

Shaders, descriptors, commands, sync, barriers

Definition at line 24 of file ComputePress.hpp.

The documentation for this class was generated from the following files:

Classes

Public Member Functions

Static Public Member Functions

Private Member Functions

Private Attributes

Static Private Attributes

Detailed Description