pulp::render::GpuDevice Class Reference

Typed handle to the device a GpuSurface owns. Copies share it. More...

#include <gpu_pass.hpp>

Public Member Functions
	GpuDevice ()=default
bool	valid () const noexcept
bool	is_lost () const
	True once Dawn reports the device lost. Non-blocking.
const GpuDeviceCapabilities &	capabilities () const
GpuResult< GpuTexture >	create_texture (const GpuTextureDesc &desc) const
GpuResult< GpuBuffer >	create_buffer (const GpuBufferDesc &desc) const
GpuResult< GpuSampler >	create_sampler (const GpuSamplerDesc &desc={}) const
GpuResult< GpuShaderModule >	create_shader_module (std::string_view wgsl, std::string_view label={}) const
	Compile WGSL.
GpuResult< GpuRenderPipeline >	create_render_pipeline (const GpuRenderPipelineDesc &desc) const
GpuResult< GpuComputePipeline >	create_compute_pipeline (const GpuComputePipelineDesc &desc) const
GpuResult< GpuBindGroup >	create_bind_group (const GpuRenderPipeline &pipeline, std::uint32_t group, const std::vector< GpuBinding > &bindings) const
	Bind group group of a pipeline's automatic layout.
GpuResult< GpuBindGroup >	create_bind_group (const GpuComputePipeline &pipeline, std::uint32_t group, const std::vector< GpuBinding > &bindings) const
GpuResult< GpuTarget >	create_target (const GpuTargetDesc &desc) const
GpuStatus	write_buffer (const GpuBuffer &buffer, std::uint64_t offset, const void *data, std::size_t size) const
	Queue a buffer write.
void	process_events () const
	Run pending Dawn callbacks once (submission status, readback).

Static Public Member Functions
static GpuResult< GpuDevice >	from_surface (GpuSurface &surface)
	Wrap the surface's device, queue and instance.

Friends
struct	detail::GpuAccess

Detailed Description

Typed handle to the device a GpuSurface owns. Copies share it.

Constructor & Destructor Documentation

GpuDevice()

pulp::render::GpuDevice::GpuDevice ( )

default

Member Function Documentation

capabilities()

const GpuDeviceCapabilities & pulp::render::GpuDevice::capabilities

(

)

const

create_bind_group() [1/2]

GpuResult< GpuBindGroup > pulp::render::GpuDevice::create_bind_group	(	const GpuComputePipeline &	pipeline,
		std::uint32_t	group,
		const std::vector< GpuBinding > &	bindings ) const

create_bind_group() [2/2]

GpuResult< GpuBindGroup > pulp::render::GpuDevice::create_bind_group	(	const GpuRenderPipeline &	pipeline,
		std::uint32_t	group,
		const std::vector< GpuBinding > &	bindings ) const

Bind group group of a pipeline's automatic layout.

create_buffer()

GpuResult< GpuBuffer > pulp::render::GpuDevice::create_buffer

(

const GpuBufferDesc &

desc

)

const

create_compute_pipeline()

GpuResult< GpuComputePipeline > pulp::render::GpuDevice::create_compute_pipeline

(

const GpuComputePipelineDesc &

desc

)

const

create_render_pipeline()

GpuResult< GpuRenderPipeline > pulp::render::GpuDevice::create_render_pipeline

(

const GpuRenderPipelineDesc &

desc

)

const

create_sampler()

GpuResult< GpuSampler > pulp::render::GpuDevice::create_sampler

(

const GpuSamplerDesc &

desc = {}

)

const

create_shader_module()

GpuResult< GpuShaderModule > pulp::render::GpuDevice::create_shader_module	(	std::string_view	wgsl,
		std::string_view	label = {} ) const

Compile WGSL.

A compile error fails with shader_compilation and the compiler's line/column diagnostics; label names the source in them.

create_target()

GpuResult< GpuTarget > pulp::render::GpuDevice::create_target

(

const GpuTargetDesc &

desc

)

const

create_texture()

GpuResult< GpuTexture > pulp::render::GpuDevice::create_texture

(

const GpuTextureDesc &

desc

)

const

from_surface()

GpuResult< GpuDevice > pulp::render::GpuDevice::from_surface ( GpuSurface & surface )

static

Wrap the surface's device, queue and instance.

Fails with unavailable before the surface is initialized, or in a build without Dawn. The surface must outlive the returned device and every object created from it.

is_lost()

bool pulp::render::GpuDevice::is_lost

(

)

const

True once Dawn reports the device lost. Non-blocking.

process_events()

void pulp::render::GpuDevice::process_events

(

)

const

Run pending Dawn callbacks once (submission status, readback).

valid()

bool pulp::render::GpuDevice::valid ( ) const

noexcept

write_buffer()

GpuStatus pulp::render::GpuDevice::write_buffer	(	const GpuBuffer &	buffer,
		std::uint64_t	offset,
		const void *	data,
		std::size_t	size ) const

Queue a buffer write.

It executes before any later submission, so write uniforms first and then submit the frame that reads them.

Friends And Related Symbol Documentation

detail::GpuAccess

friend struct detail::GpuAccess

friend

---

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

• /home/runner/work/pulp/pulp/core/render/include/pulp/render/gpu_pass.hpp