Files
test2/source/blender/gpu/vulkan/vk_command_buffer.hh

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

237 lines
8.0 KiB
C++
Raw Normal View History

/* SPDX-FileCopyrightText: 2023 Blender Foundation
*
* SPDX-License-Identifier: GPL-2.0-or-later */
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
/** \file
* \ingroup gpu
*/
#pragma once
#include "vk_common.hh"
#include "vk_resource_tracker.hh"
#include "BLI_utility_mixins.hh"
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
namespace blender::gpu {
class VKBuffer;
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
struct VKBufferWithOffset;
class VKDescriptorSet;
class VKFrameBuffer;
class VKIndexBuffer;
class VKPipeline;
class VKPushConstants;
class VKStorageBuffer;
class VKTexture;
class VKVertexBuffer;
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
2023-02-27 21:44:59 +11:00
/** Command buffer to keep track of the life-time of a command buffer. */
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
class VKCommandBuffer : NonCopyable, NonMovable {
/** Not owning handle to the command buffer and device. Handle is owned by `GHOST_ContextVK`. */
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
VkDevice vk_device_ = VK_NULL_HANDLE;
VkCommandBuffer vk_command_buffer_ = VK_NULL_HANDLE;
VkQueue vk_queue_ = VK_NULL_HANDLE;
/**
* Timeout to use when waiting for fences in nanoseconds.
*
* Currently added as the fence will halt when there are no commands in the command buffer for
* the second time. This should be solved and this timeout should be removed.
*/
static constexpr uint64_t FenceTimeout = UINT64_MAX;
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
/** Owning handles */
VkFence vk_fence_ = VK_NULL_HANDLE;
VKSubmissionID submission_id_;
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
private:
enum class Stage {
Initial,
Recording,
BetweenRecordingAndSubmitting,
Submitted,
Executed,
};
/*
* Some vulkan command require an active frame buffer. Others require no active frame-buffer. As
* our current API does not provide a solution for this we need to keep track of the actual state
* and do the changes when recording the next command.
*
* This is a temporary solution to get things rolling.
* TODO: In a future solution we should decide the scope of a command buffer.
*
* - command buffer per draw command.
* - minimize command buffers and track render passes.
* - add custom encoder to also track resource usages.
*
2023-04-26 12:01:23 +10:00
* Currently I expect the custom encoder has to be done eventually. But want to keep postponing
* the custom encoder for now to collect more use cases it should solve. (first pixel drawn on
* screen).
*
* Some command can also be encoded in another way when encoded as a first command. For example
2023-04-26 12:01:23 +10:00
* clearing a frame-buffer textures isn't allowed inside a render pass, but clearing the
* frame-buffer textures via ops is allowed. When clearing a frame-buffer texture directly after
* beginning a render pass could be re-encoded to do this in the same command.
*
* So for now we track the state and temporary switch to another state if the command requires
* it.
*/
struct {
/* Reference to the last_framebuffer where begin_render_pass was called for. */
const VKFrameBuffer *framebuffer_ = nullptr;
/* Is last_framebuffer_ currently bound. Each call should ensure the correct state. */
bool framebuffer_active_ = false;
/* Amount of times a check has been requested. */
uint64_t checks_ = 0;
/* Amount of times a check required to change the render pass. */
uint64_t switches_ = 0;
/* Number of times a vkDraw command has been recorded. */
uint64_t draw_counts = 0;
/**
* Current stage of the command buffer to keep track of inconsistencies & incorrect usage.
*/
Stage stage = Stage::Initial;
} state;
bool is_in_stage(Stage stage)
{
return state.stage == stage;
}
void stage_set(Stage stage)
{
state.stage = stage;
}
std::string to_string(Stage stage)
{
switch (stage) {
case Stage::Initial:
return "INITIAL";
case Stage::Recording:
return "RECORDING";
case Stage::BetweenRecordingAndSubmitting:
return "BEFORE_SUBMIT";
case Stage::Submitted:
return "SUBMITTED";
case Stage::Executed:
return "EXECUTED";
}
return "UNKNOWN";
}
void stage_transfer(Stage stage_from, Stage stage_to)
{
BLI_assert(is_in_stage(stage_from));
UNUSED_VARS_NDEBUG(stage_from);
#if 0
printf(" *** Transfer stage from %s to %s\n",
to_string(stage_from).c_str(),
to_string(stage_to).c_str());
#endif
stage_set(stage_to);
}
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
public:
virtual ~VKCommandBuffer();
void init(const VkDevice vk_device, const VkQueue vk_queue, VkCommandBuffer vk_command_buffer);
void begin_recording();
void end_recording();
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
void bind(const VKPipeline &vk_pipeline, VkPipelineBindPoint bind_point);
void bind(const VKDescriptorSet &descriptor_set,
const VkPipelineLayout vk_pipeline_layout,
VkPipelineBindPoint bind_point);
void bind(const uint32_t binding,
const VKVertexBuffer &vertex_buffer,
const VkDeviceSize offset);
/* Bind the given buffer as a vertex buffer. */
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
void bind(const uint32_t binding, const VKBufferWithOffset &vertex_buffer);
void bind(const uint32_t binding, const VkBuffer &vk_vertex_buffer, const VkDeviceSize offset);
/* Bind the given buffer as an index buffer. */
void bind(const VKBufferWithOffset &index_buffer, VkIndexType index_type);
void begin_render_pass(const VKFrameBuffer &framebuffer);
void end_render_pass(const VKFrameBuffer &framebuffer);
/**
* Add a push constant command to the command buffer.
*
* Only valid when the storage type of push_constants is StorageType::PUSH_CONSTANTS.
*/
void push_constants(const VKPushConstants &push_constants,
const VkPipelineLayout vk_pipeline_layout,
const VkShaderStageFlags vk_shader_stages);
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
void dispatch(int groups_x_len, int groups_y_len, int groups_z_len);
void dispatch(VKStorageBuffer &command_buffer);
2023-02-27 20:54:27 +11:00
/** Copy the contents of a texture MIP level to the dst buffer. */
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
void copy(VKBuffer &dst_buffer, VKTexture &src_texture, Span<VkBufferImageCopy> regions);
Vulkan: Texture Data Conversions This PR adds basic support for texture update, read back and clearing for Vulkan. In Vulkan we need to convert each data type ourselves as vulkan buffers are untyped. Therefore this change mostly is about data conversions. Considerations: - Use a compute shader to do the conversions: - Leads to performance regression as compute pipeline can stall graphics pipeline - Lead to additional memory usage as two staging buffers are needed one to hold the CPU data, and one to hold the converted data. - Do inline conversion when sending the data to Vulkan using `eGPUDataFormat` - Additional CPU cycles required and not easy to optimize as it the implementation requires many branches. - Do inline conversion when sending the data to Vulkan (optimized for CPU) For this solution it was chosen to implement the 3rd option as it is fast and doesn't require additional memory what the other options do. **Use Imath/half.h** This patch uses `Imath/half.h` (dependency of OpenEXR) similar to alembic. But this makes vulkan dependent of the availability of OpenEXR. For now this isn't checked, but when we are closer to a working Vulkan backend we have to make a decision how to cope with this dependency. **Missing Features** *Framebuffer textures* This doesn't include all possible data transformations. Some of those transformation can only be tested after the VKFramebuffer has been implemented. Some texture types are only available when created for a framebuffer. These include the depth and stencil variations. *Component format* Is more relevant when implementing VKVertexBuffer. *SRGB textures* SRGB encoded textures aren't natively supported on all platforms, in all usages and might require workarounds. This should be done in a separate PR in a later stage when we are required to use SRGB textures. **Test cases** The added test cases gives an overview of the missing bits and pieces of the patch. When the implementation/direction is accepted more test cases can be enabled/implemented. Some of these test cases will skip depending on the actual support of platform the tests are running on. For example OpenGL/NVidia will skip the next test as it doesn't support the texture format on OpenGL, although it does support it on Vulkan. ``` [ RUN ] GPUOpenGLTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ SKIPPED ] GPUOpenGLTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ RUN ] GPUVulkanTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ OK ] GPUVulkanTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI ``` Pull Request: https://projects.blender.org/blender/blender/pulls/105762
2023-03-24 08:09:19 +01:00
void copy(VKTexture &dst_texture, VKBuffer &src_buffer, Span<VkBufferImageCopy> regions);
void copy(VKTexture &dst_texture, VKTexture &src_texture, Span<VkImageCopy> regions);
void blit(VKTexture &dst_texture, VKTexture &src_texture, Span<VkImageBlit> regions);
void blit(VKTexture &dst_texture,
VkImageLayout dst_layout,
VKTexture &src_texture,
VkImageLayout src_layout,
Span<VkImageBlit> regions);
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
void pipeline_barrier(VkPipelineStageFlags source_stages,
VkPipelineStageFlags destination_stages);
void pipeline_barrier(Span<VkImageMemoryBarrier> image_memory_barriers);
Vulkan: Texture Data Conversions This PR adds basic support for texture update, read back and clearing for Vulkan. In Vulkan we need to convert each data type ourselves as vulkan buffers are untyped. Therefore this change mostly is about data conversions. Considerations: - Use a compute shader to do the conversions: - Leads to performance regression as compute pipeline can stall graphics pipeline - Lead to additional memory usage as two staging buffers are needed one to hold the CPU data, and one to hold the converted data. - Do inline conversion when sending the data to Vulkan using `eGPUDataFormat` - Additional CPU cycles required and not easy to optimize as it the implementation requires many branches. - Do inline conversion when sending the data to Vulkan (optimized for CPU) For this solution it was chosen to implement the 3rd option as it is fast and doesn't require additional memory what the other options do. **Use Imath/half.h** This patch uses `Imath/half.h` (dependency of OpenEXR) similar to alembic. But this makes vulkan dependent of the availability of OpenEXR. For now this isn't checked, but when we are closer to a working Vulkan backend we have to make a decision how to cope with this dependency. **Missing Features** *Framebuffer textures* This doesn't include all possible data transformations. Some of those transformation can only be tested after the VKFramebuffer has been implemented. Some texture types are only available when created for a framebuffer. These include the depth and stencil variations. *Component format* Is more relevant when implementing VKVertexBuffer. *SRGB textures* SRGB encoded textures aren't natively supported on all platforms, in all usages and might require workarounds. This should be done in a separate PR in a later stage when we are required to use SRGB textures. **Test cases** The added test cases gives an overview of the missing bits and pieces of the patch. When the implementation/direction is accepted more test cases can be enabled/implemented. Some of these test cases will skip depending on the actual support of platform the tests are running on. For example OpenGL/NVidia will skip the next test as it doesn't support the texture format on OpenGL, although it does support it on Vulkan. ``` [ RUN ] GPUOpenGLTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ SKIPPED ] GPUOpenGLTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ RUN ] GPUVulkanTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI [ OK ] GPUVulkanTest.texture_roundtrip__GPU_DATA_2_10_10_10_REV__GPU_RGB10_A2UI ``` Pull Request: https://projects.blender.org/blender/blender/pulls/105762
2023-03-24 08:09:19 +01:00
/**
* Clear color image resource.
*/
void clear(VkImage vk_image,
VkImageLayout vk_image_layout,
const VkClearColorValue &vk_clear_color,
Span<VkImageSubresourceRange> ranges);
/**
* Clear attachments of the active framebuffer.
*/
void clear(Span<VkClearAttachment> attachments, Span<VkClearRect> areas);
void fill(VKBuffer &buffer, uint32_t data);
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
void draw(int v_first, int v_count, int i_first, int i_count);
void draw(
int index_count, int instance_count, int first_index, int vertex_offset, int first_instance);
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
/**
* Stop recording commands, encode + send the recordings to Vulkan, wait for the until the
* commands have been executed and start the command buffer to accept recordings again.
*/
void submit();
const VKSubmissionID &submission_id_get() const
{
return submission_id_;
}
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
private:
void encode_recorded_commands();
void submit_encoded_commands();
/**
* Validate that there isn't a framebuffer being tracked (bound or not bound).
*
* Raises an assert in debug when a framebuffer is being tracked.
*/
void validate_framebuffer_not_exists();
/**
* Validate that there is a framebuffer being tracked (bound or not bound).
*
* Raises an assert in debug when no framebuffer is being tracked.
*/
void validate_framebuffer_exists();
/**
* Ensure that there is no framebuffer being tracked or the tracked framebuffer isn't bound.
*/
void ensure_no_active_framebuffer();
/**
* Ensure that the tracked framebuffer is bound.
*/
void ensure_active_framebuffer();
Vulkan: Initial Compute Shaders support This patch adds initial support for compute shaders to the vulkan backend. As the development is oriented to the test- cases we have the implementation is limited to what is used there. It has been validated that with this patch that the following test cases are running as expected - `GPUVulkanTest.gpu_shader_compute_vbo` - `GPUVulkanTest.gpu_shader_compute_ibo` - `GPUVulkanTest.gpu_shader_compute_ssbo` - `GPUVulkanTest.gpu_storage_buffer_create_update_read` - `GPUVulkanTest.gpu_shader_compute_2d` This patch includes: - Allocating VkBuffer on device. - Uploading data from CPU to VkBuffer. - Binding VkBuffer as SSBO to a compute shader. - Execute compute shader and altering VkBuffer. - Download the VkBuffer to CPU ram. - Validate that it worked. - Use device only vertex buffer as SSBO - Use device only index buffer as SSBO - Use device only image buffers GHOST API has been changed as the original design was created before we even had support for compute shaders in blender. The function `GHOST_getVulkanBackbuffer` has been separated to retrieve the command buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order to do correct command buffer processing we needed access to the queue owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles` function. Open topics (not considered part of this patch) - Memory barriers & command buffer encoding - Indirect compute dispatching - Rest of the test cases - Data conversions when requested data format is different than on device. - GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices. NVIDIA doesn't seem to support 1d textures. Pull-request: #104518
2023-02-21 15:03:12 +01:00
};
} // namespace blender::gpu