Files
test/source/blender/gpu/vulkan/vk_descriptor_set.hh
Jeroen Bakker 809a5aa418 Vulkan: Initial Graphics Pipeline
Initial graphic pipeline targeting. The goal of this PR is to have an initial
graphics pipeline with missing features. It should help identifying
areas that requires engineering.

Current state is that developers of the GPU module can help with the many
smaller pieces that needs to be engineered in order to get it working. It is not
intended for users or developers from other modules, but your welcome to learn
and give feedback on the code and engineering part.

We do expect that large parts of the code still needs to be re-engineered into
a more future-proof implementation.

**Some highlights**:
- In Vulkan the state is kept in the pipeline. Therefore the state is tracked
  per pipeline. In the near future this could be used as a cache. More research
  is needed against the default pipeline cache that vulkan already provides.
- This PR is based on the work that Kazashi Yoshioka already did. And include
  work from him in the next areas
  - Vertex attributes
  - Vertex data conversions
  - Pipeline state
- Immediate support working.
- This PR modifies the VKCommandBuffer to keep track of the framebuffer and its
  binding state(render pass). Some Vulkan commands require no render pass to be
  active, other require a render pass. As the order of our commands on API level
  can not be separated this PR introduces a state engine to keep track of the
  current state and desired state. This is a temporary solution, the final
  solution will be proposed when we have a pixel on the screen. At that time
  I expect that we can design a command encoder that supports all the cases
  we need.

**Notices**:
- This branch works on NVIDIA GPUs and has been validated on a Linux system. AMD
  is known not to work (stalls) and Intel GPUs have not been tested at all. Windows might work
  but hasn't been validated yet.
- The graphics pipeline is implemented with pixels in mind, not with performance. Currently
  when a draw call is scheduled it is flushed and waited until it is finished drawing, before
  other draw calls can be scheduled. We expected the performance to be worse that it actually
  is, but we expect huge performance gains in the future.
- Any advanced drawing (that is used by the image editor, compositor or 3d viewport) isn't
  implemented and might crash when used.
- Using multiple windows or resizing of window isn't supported and will stall the system.

Pull Request: https://projects.blender.org/blender/blender/pulls/106224
2023-05-11 13:01:56 +02:00

190 lines
4.9 KiB
C++

/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation */
/** \file
* \ingroup gpu
*/
#pragma once
#include "BLI_utility_mixins.hh"
#include "BLI_vector.hh"
#include "gpu_shader_private.hh"
#include "vk_buffer.hh"
#include "vk_common.hh"
#include "vk_resource_tracker.hh"
#include "vk_uniform_buffer.hh"
namespace blender::gpu {
class VKIndexBuffer;
class VKShaderInterface;
class VKStorageBuffer;
class VKTexture;
class VKUniformBuffer;
class VKVertexBuffer;
class VKDescriptorSetTracker;
class VKSampler;
/**
* In vulkan shader resources (images and buffers) are grouped in descriptor sets.
*
* The resources inside a descriptor set can be updated and bound per set.
*
* Currently Blender only supports a single descriptor set per shader, but it is planned to be able
* to use 2 descriptor sets per shader. One for each #blender::gpu::shader::Frequency.
*/
class VKDescriptorSet : NonCopyable {
public:
/**
* Binding location of a resource in a descriptor set.
*
* Locations and bindings are used for different reasons. In the Vulkan backend we use
* ShaderInput.location to store the descriptor set + the resource binding inside the descriptor
* set. To ease the development the VKDescriptorSet::Location will be used to hide this
* confusion.
*
* NOTE: [future development] When supporting multiple descriptor sets the encoding/decoding can
* be centralized here. Location will then also contain the descriptor set index.
*/
struct Location {
private:
/**
* References to a binding in the descriptor set.
*/
uint32_t binding;
Location(uint32_t binding) : binding(binding) {}
public:
Location() = default;
bool operator==(const Location &other) const
{
return binding == other.binding;
}
operator uint32_t() const
{
return binding;
}
friend class VKDescriptorSetTracker;
friend class VKShaderInterface;
};
VkDescriptorPool vk_descriptor_pool_ = VK_NULL_HANDLE;
VkDescriptorSet vk_descriptor_set_ = VK_NULL_HANDLE;
public:
VKDescriptorSet() = default;
VKDescriptorSet(VkDescriptorPool vk_descriptor_pool, VkDescriptorSet vk_descriptor_set)
: vk_descriptor_pool_(vk_descriptor_pool), vk_descriptor_set_(vk_descriptor_set)
{
}
VKDescriptorSet(VKDescriptorSet &&other);
virtual ~VKDescriptorSet();
VKDescriptorSet &operator=(VKDescriptorSet &&other)
{
vk_descriptor_set_ = other.vk_descriptor_set_;
vk_descriptor_pool_ = other.vk_descriptor_pool_;
other.mark_freed();
return *this;
}
VkDescriptorSet vk_handle() const
{
return vk_descriptor_set_;
}
VkDescriptorPool vk_pool_handle() const
{
return vk_descriptor_pool_;
}
void mark_freed();
};
class VKDescriptorSetTracker : protected VKResourceTracker<VKDescriptorSet> {
friend class VKDescriptorSet;
public:
struct Binding {
VKDescriptorSet::Location location;
VkDescriptorType type;
VkBuffer vk_buffer = VK_NULL_HANDLE;
VkDeviceSize buffer_size = 0;
VKTexture *texture = nullptr;
VkSampler vk_sampler = VK_NULL_HANDLE;
Binding()
{
location.binding = 0;
}
bool is_buffer() const
{
return ELEM(type, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER);
}
bool is_image() const
{
return ELEM(type,
VK_DESCRIPTOR_TYPE_STORAGE_IMAGE,
VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER) &&
texture != nullptr;
}
};
private:
/** A list of bindings that needs to be updated. */
Vector<Binding> bindings_;
VkDescriptorSetLayout layout_ = VK_NULL_HANDLE;
public:
VKDescriptorSetTracker() {}
VKDescriptorSetTracker(VkDescriptorSetLayout layout) : layout_(layout) {}
void bind_as_ssbo(VKVertexBuffer &buffer, VKDescriptorSet::Location location);
void bind_as_ssbo(VKIndexBuffer &buffer, VKDescriptorSet::Location location);
void bind(VKStorageBuffer &buffer, VKDescriptorSet::Location location);
void bind(VKUniformBuffer &buffer, VKDescriptorSet::Location location);
/* TODO: bind as image */
void image_bind(VKTexture &texture, VKDescriptorSet::Location location);
void bind(VKTexture &texture, VKDescriptorSet::Location location, VKSampler &sampler);
/**
* Some shaders don't need any descriptor sets so we don't need to bind them.
*
* The result of this function determines if the descriptor set has any layout assigned.
* TODO: we might want to make descriptor sets optional for pipelines.
*/
bool has_layout() const
{
return layout_ != VK_NULL_HANDLE;
}
/**
* Update the descriptor set on the device.
*/
void update(VKContext &context);
std::unique_ptr<VKDescriptorSet> &active_descriptor_set()
{
return active_resource();
}
protected:
std::unique_ptr<VKDescriptorSet> create_resource(VKContext &context) override;
private:
Binding &ensure_location(VKDescriptorSet::Location location);
};
} // namespace blender::gpu