tmaq
/
tornavis
1
0
Fork 0
Code Issues 2 Pull Requests 29 Packages Projects 24 Releases Wiki Activity

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
This commit is contained in:
Jeroen Bakker 2023-02-21 15:03:12 +01:00
parent 02c2970983
commit 7fb1f060ff
49 changed files with 2477 additions and 156 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

83
intern/ghost/GHOST_C-api.h
View File

@ -1195,24 +1195,89 @@ int GHOST_XrGetControllerModelData(GHOST_XrContextHandle xr_context,
#ifdef WITH_VULKAN_BACKEND
/**
* Return VULKAN handles for the given context.
* Get Vulkan handles for the given context.
*
* These handles are the same for a given context.
* Should should only be called when using a Vulkan context.
* Other contexts will not return any handles and leave the
* handles where the parameters are referring to unmodified.
*
* \param context: GHOST context handle of a vulkan context to
* get the Vulkan handles from.
* \param r_instance: After calling this function the VkInstance
* referenced by this parameter will contain the VKInstance handle
* of the context associated with the `context` parameter.
* \param r_physical_device: After calling this function the VkPhysicalDevice
* referenced by this parameter will contain the VKPhysicalDevice handle
* of the context associated with the `context` parameter.
* \param r_device: After calling this function the VkDevice
* referenced by this parameter will contain the VKDevice handle
* of the context associated with the `context` parameter.
* \param r_graphic_queue_family: After calling this function the uint32_t
* referenced by this parameter will contain the graphic queue family id
* of the context associated with the `context` parameter.
* \param r_queue: After calling this function the VkQueue
* referenced by this parameter will contain the VKQueue handle
* of the context associated with the `context` parameter.
*/
void GHOST_GetVulkanHandles(GHOST_ContextHandle context,
void *r_instance,
void *r_physical_device,
void *r_device,
uint32_t *r_graphic_queue_family);
uint32_t *r_graphic_queue_family,
void *r_queue);
/**
* Return VULKAN back-buffer resources handles for the given window.
* Return Vulkan command buffer.
*
* Command buffers are different for each image in the swap chain.
* At the start of each frame the correct command buffer should be
* retrieved with this function.
*
* Should should only be called when using a Vulkan context.
* Other contexts will not return any handles and leave the
* handles where the parameters are referring to unmodified.
*
* \param context: GHOST context handle to a vulkan context to get the
* command queue from.
* \param r_command_buffer: After calling this function the VkCommandBuffer
* referenced by this parameter will contain the VKCommandBuffer handle
* of the current back buffer (when swap chains are enabled) or
* it will contain a general VkCommandQueue.
*/
void GHOST_GetVulkanCommandBuffer(GHOST_ContextHandle context, void *r_command_buffer);
/**
* Gets the Vulkan backbuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap event as the backbuffer will change.
*
* Should should only be called when using a Vulkan context with an active swap chain.
* Other contexts will not return any handles and leave the
* handles where the parameters are referring to unmodified.
*
* \param windowhandle: GHOST window handle to a window to get the resource from.
* \param r_image: After calling this function the VkImage
* referenced by this parameter will contain the VKImage handle
* of the current back buffer.
* \param r_framebuffer: After calling this function the VkFramebuffer
* referenced by this parameter will contain the VKFramebuffer handle
* of the current back buffer.
* \param r_render_pass: After calling this function the VkRenderPass
* referenced by this parameter will contain the VKRenderPass handle
* of the current back buffer.
* \param r_extent: After calling this function the VkExtent2D
* referenced by this parameter will contain the size of the
* frame buffer and image in pixels.
* \param r_fb_id: After calling this function the uint32_t
* referenced by this parameter will contain the id of the
* framebuffer of the current back buffer.
*/
void GHOST_GetVulkanBackbuffer(GHOST_WindowHandle windowhandle,
void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id);
void *r_image,
void *r_framebuffer,
void *r_render_pass,
void *r_extent,
uint32_t *r_fb_id);
#endif

85
intern/ghost/GHOST_IContext.h
View File

@ -40,19 +40,84 @@ class GHOST_IContext {
virtual unsigned int getDefaultFramebuffer() = 0;
virtual GHOST_TSuccess getVulkanHandles(void *, void *, void *, uint32_t *) = 0;
/**
* Get Vulkan handles for the given context.
*
* These handles are the same for a given context.
* Should should only be called when using a Vulkan context.
* Other contexts will not return any handles and leave the
* handles where the parameters are referring to unmodified.
*
* \param r_instance: After calling this function the VkInstance
* referenced by this parameter will contain the VKInstance handle
* of the context associated with the `context` parameter.
* \param r_physical_device: After calling this function the VkPhysicalDevice
* referenced by this parameter will contain the VKPhysicalDevice handle
* of the context associated with the `context` parameter.
* \param r_device: After calling this function the VkDevice
* referenced by this parameter will contain the VKDevice handle
* of the context associated with the `context` parameter.
* \param r_graphic_queue_family: After calling this function the uint32_t
* referenced by this parameter will contain the graphic queue family id
* of the context associated with the `context` parameter.
* \param r_queue: After calling this function the VkQueue
* referenced by this parameter will contain the VKQueue handle
* of the context associated with the `context` parameter.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanHandles(void *r_instance,
void *r_physical_device,
void *r_device,
uint32_t *r_graphic_queue_family,
void *r_queue) = 0;
/**
* Gets the Vulkan framebuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap events as the framebuffer will change.
* \return A boolean success indicator.
* Return Vulkan command buffer.
*
* Command buffers are different for each image in the swap chain.
* At the start of each frame the correct command buffer should be
* retrieved with this function.
*
* \param r_command_buffer: After calling this function the VkCommandBuffer
* referenced by this parameter will contain the VKCommandBuffer handle
* of the current back buffer (when swap chains are enabled) or
* it will contain a general VkCommandQueue.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id) = 0;
virtual GHOST_TSuccess getVulkanCommandBuffer(void *r_command_buffer) = 0;
/**
* Gets the Vulkan backbuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap event as the backbuffer will change.
*
* \param r_image: After calling this function the VkImage
* referenced by this parameter will contain the VKImage handle
* of the current back buffer.
* \param r_framebuffer: After calling this function the VkFramebuffer
* referenced by this parameter will contain the VKFramebuffer handle
* of the current back buffer.
* \param r_render_pass: After calling this function the VkRenderPass
* referenced by this parameter will contain the VKRenderPass handle
* of the current back buffer.
* \param r_extent: After calling this function the VkExtent2D
* referenced by this parameter will contain the size of the
* frame buffer and image in pixels.
* \param r_fb_id: After calling this function the uint32_t
* referenced by this parameter will contain the id of the
* framebuffer of the current back buffer.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void *r_image,
void *r_framebuffer,
void *r_render_pass,
void *r_extent,
uint32_t *r_fb_id) = 0;
virtual GHOST_TSuccess swapBuffers() = 0;

1
intern/ghost/GHOST_IWindow.h
View File

@ -217,7 +217,6 @@ class GHOST_IWindow {
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id) = 0;

15
intern/ghost/intern/GHOST_C-api.cpp
View File

@ -1203,22 +1203,29 @@ void GHOST_GetVulkanHandles(GHOST_ContextHandle contexthandle,
void *r_instance,
void *r_physical_device,
void *r_device,
uint32_t *r_graphic_queue_family)
uint32_t *r_graphic_queue_family,
void *r_queue)
{
GHOST_IContext *context = (GHOST_IContext *)contexthandle;
context->getVulkanHandles(r_instance, r_physical_device, r_device, r_graphic_queue_family);
context->getVulkanHandles(
r_instance, r_physical_device, r_device, r_graphic_queue_family, r_queue);
}
void GHOST_GetVulkanCommandBuffer(GHOST_ContextHandle contexthandle, void *r_command_buffer)
{
GHOST_IContext *context = (GHOST_IContext *)contexthandle;
context->getVulkanCommandBuffer(r_command_buffer);
}
void GHOST_GetVulkanBackbuffer(GHOST_WindowHandle windowhandle,
void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id)
{
GHOST_IWindow *window = (GHOST_IWindow *)windowhandle;
window->getVulkanBackbuffer(image, framebuffer, command_buffer, render_pass, extent, fb_id);
window->getVulkanBackbuffer(image, framebuffer, render_pass, extent, fb_id);
}
#endif /* WITH_VULKAN */

83
intern/ghost/intern/GHOST_Context.h
View File

@ -136,27 +136,88 @@ class GHOST_Context : public GHOST_IContext {
}
/**
* Gets the Vulkan context related resource handles.
* \return A boolean success indicator.
* Get Vulkan handles for the given context.
*
* These handles are the same for a given context.
* Should should only be called when using a Vulkan context.
* Other contexts will not return any handles and leave the
* handles where the parameters are referring to unmodified.
*
* \param r_instance: After calling this function the VkInstance
* referenced by this parameter will contain the VKInstance handle
* of the context associated with the `context` parameter.
* \param r_physical_device: After calling this function the VkPhysicalDevice
* referenced by this parameter will contain the VKPhysicalDevice handle
* of the context associated with the `context` parameter.
* \param r_device: After calling this function the VkDevice
* referenced by this parameter will contain the VKDevice handle
* of the context associated with the `context` parameter.
* \param r_graphic_queue_family: After calling this function the uint32_t
* referenced by this parameter will contain the graphic queue family id
* of the context associated with the `context` parameter.
* \param r_queue: After calling this function the VkQueue
* referenced by this parameter will contain the VKQueue handle
* of the context associated with the `context` parameter.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanHandles(void * /*r_instance*/,
void * /*r_physical_device*/,
void * /*r_device*/,
uint32_t * /*r_graphic_queue_family*/) override
uint32_t * /*r_graphic_queue_family*/,
void * /*r_queue*/) override
{
return GHOST_kFailure;
};
/**
* Gets the Vulkan framebuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap events as the framebuffer will change.
* \return A boolean success indicator.
* Return Vulkan command buffer.
*
* Command buffers are different for each image in the swap chain.
* At the start of each frame the correct command buffer should be
* retrieved with this function.
*
* \param r_command_buffer: After calling this function the VkCommandBuffer
* referenced by this parameter will contain the VKCommandBuffer handle
* of the current back buffer (when swap chains are enabled) or
* it will contain a general VkCommandQueue.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void * /*image*/,
void * /*framebuffer*/,
void * /*command_buffer*/,
void * /*render_pass*/,
void * /*extent*/,
virtual GHOST_TSuccess getVulkanCommandBuffer(void * /*r_command_buffer*/) override
{
return GHOST_kFailure;
};
/**
* Gets the Vulkan backbuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap event as the backbuffer will change.
*
* \param r_image: After calling this function the VkImage
* referenced by this parameter will contain the VKImage handle
* of the current back buffer.
* \param r_framebuffer: After calling this function the VkFramebuffer
* referenced by this parameter will contain the VKFramebuffer handle
* of the current back buffer.
* \param r_render_pass: After calling this function the VkRenderPass
* referenced by this parameter will contain the VKRenderPass handle
* of the current back buffer.
* \param r_extent: After calling this function the VkExtent2D
* referenced by this parameter will contain the size of the
* frame buffer and image in pixels.
* \param r_fb_id: After calling this function the uint32_t
* referenced by this parameter will contain the id of the
* framebuffer of the current back buffer.
* \returns GHOST_kFailure when context isn't a Vulkan context.
* GHOST_kSuccess when the context is a Vulkan context and the
* handles have been set.
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void * /*r_image*/,
void * /*r_framebuffer*/,
void * /*r_render_pass*/,
void * /*r_extent*/,
uint32_t * /*fb_id*/) override
{
return GHOST_kFailure;

82
intern/ghost/intern/GHOST_ContextVK.cpp
View File

@ -288,19 +288,14 @@ GHOST_TSuccess GHOST_ContextVK::swapBuffers()
return GHOST_kSuccess;
}
GHOST_TSuccess GHOST_ContextVK::getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id)
GHOST_TSuccess GHOST_ContextVK::getVulkanBackbuffer(
void *image, void *framebuffer, void *render_pass, void *extent, uint32_t *fb_id)
{
if (m_swapchain == VK_NULL_HANDLE) {
return GHOST_kFailure;
}
*((VkImage *)image) = m_swapchain_images[m_currentImage];
*((VkFramebuffer *)framebuffer) = m_swapchain_framebuffers[m_currentImage];
*((VkCommandBuffer *)command_buffer) = m_command_buffers[m_currentImage];
*((VkRenderPass *)render_pass) = m_render_pass;
*((VkExtent2D *)extent) = m_render_extent;
*fb_id = m_swapchain_id * 10 + m_currentFrame;
@ -311,12 +306,30 @@ GHOST_TSuccess GHOST_ContextVK::getVulkanBackbuffer(void *image,
GHOST_TSuccess GHOST_ContextVK::getVulkanHandles(void *r_instance,
void *r_physical_device,
void *r_device,
uint32_t *r_graphic_queue_family)
uint32_t *r_graphic_queue_family,
void *r_queue)
{
*((VkInstance *)r_instance) = m_instance;
*((VkPhysicalDevice *)r_physical_device) = m_physical_device;
*((VkDevice *)r_device) = m_device;
*r_graphic_queue_family = m_queue_family_graphic;
*((VkQueue *)r_queue) = m_graphic_queue;
return GHOST_kSuccess;
}
GHOST_TSuccess GHOST_ContextVK::getVulkanCommandBuffer(void *r_command_buffer)
{
if (m_command_buffers.empty()) {
return GHOST_kFailure;
}
if (m_swapchain == VK_NULL_HANDLE) {
*((VkCommandBuffer *)r_command_buffer) = m_command_buffers[0];
}
else {
*((VkCommandBuffer *)r_command_buffer) = m_command_buffers[m_currentImage];
}
return GHOST_kSuccess;
}
@ -520,6 +533,9 @@ static GHOST_TSuccess getGraphicQueueFamily(VkPhysicalDevice device, uint32_t *r
*r_queue_index = 0;
for (const auto &queue_family : queue_families) {
/* Every vulkan implementation by spec must have one queue family that support both graphics
* and compute pipelines. We select this one; compute only queue family hints at async compute
* implementations.*/
if ((queue_family.queueFlags & VK_QUEUE_GRAPHICS_BIT) &&
(queue_family.queueFlags & VK_QUEUE_COMPUTE_BIT)) {
return GHOST_kSuccess;
@ -619,16 +635,36 @@ static GHOST_TSuccess selectPresentMode(VkPhysicalDevice device,
return GHOST_kFailure;
}
GHOST_TSuccess GHOST_ContextVK::createCommandBuffers()
GHOST_TSuccess GHOST_ContextVK::createCommandPools()
{
m_command_buffers.resize(m_swapchain_image_views.size());
VkCommandPoolCreateInfo poolInfo = {};
poolInfo.sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO;
poolInfo.flags = VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT;
poolInfo.queueFamilyIndex = m_queue_family_graphic;
VK_CHECK(vkCreateCommandPool(m_device, &poolInfo, NULL, &m_command_pool));
return GHOST_kSuccess;
}
GHOST_TSuccess GHOST_ContextVK::createGraphicsCommandBuffer()
{
assert(m_command_pool != VK_NULL_HANDLE);
assert(m_command_buffers.size() == 0);
m_command_buffers.resize(1);
VkCommandBufferAllocateInfo alloc_info = {};
alloc_info.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO;
alloc_info.commandPool = m_command_pool;
alloc_info.level = VK_COMMAND_BUFFER_LEVEL_PRIMARY;
alloc_info.commandBufferCount = static_cast<uint32_t>(m_command_buffers.size());
VK_CHECK(vkAllocateCommandBuffers(m_device, &alloc_info, m_command_buffers.data()));
return GHOST_kSuccess;
}
GHOST_TSuccess GHOST_ContextVK::createGraphicsCommandBuffers()
{
assert(m_command_pool != VK_NULL_HANDLE);
m_command_buffers.resize(m_swapchain_image_views.size());
VkCommandBufferAllocateInfo alloc_info = {};
alloc_info.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO;
@ -637,7 +673,6 @@ GHOST_TSuccess GHOST_ContextVK::createCommandBuffers()
alloc_info.commandBufferCount = static_cast<uint32_t>(m_command_buffers.size());
VK_CHECK(vkAllocateCommandBuffers(m_device, &alloc_info, m_command_buffers.data()));
return GHOST_kSuccess;
}
@ -776,7 +811,7 @@ GHOST_TSuccess GHOST_ContextVK::createSwapchain()
VK_CHECK(vkCreateFence(m_device, &fence_info, NULL, &m_in_flight_fences[i]));
}
createCommandBuffers();
createGraphicsCommandBuffers();
return GHOST_kSuccess;
}
@ -841,6 +876,13 @@ GHOST_TSuccess GHOST_ContextVK::initializeDrawingContext()
extensions_device.push_back(VK_KHR_SWAPCHAIN_EXTENSION_NAME);
}
extensions_device.push_back("VK_KHR_dedicated_allocation");
extensions_device.push_back("VK_KHR_get_memory_requirements2");
/* Enable MoltenVK required instance extensions.*/
#ifdef VK_MVK_MOLTENVK_EXTENSION_NAME
requireExtension(
extensions_available, extensions_enabled, "VK_KHR_get_physical_device_properties2");
#endif
VkApplicationInfo app_info = {};
app_info.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO;
@ -903,6 +945,15 @@ GHOST_TSuccess GHOST_ContextVK::initializeDrawingContext()
return GHOST_kFailure;
}
#ifdef VK_MVK_MOLTENVK_EXTENSION_NAME
/* According to the Vulkan specs, when `VK_KHR_portability_subset` is available it should be
* enabled. See
* https://vulkan.lunarg.com/doc/view/1.2.198.1/mac/1.2-extensions/vkspec.html#VUID-VkDeviceCreateInfo-pProperties-04451*/
if (device_extensions_support(m_physical_device, {VK_KHR_PORTABILITY_SUBSET_EXTENSION_NAME})) {
extensions_device.push_back(VK_KHR_PORTABILITY_SUBSET_EXTENSION_NAME);
}
#endif
vector<VkDeviceQueueCreateInfo> queue_create_infos;
{
@ -962,11 +1013,14 @@ GHOST_TSuccess GHOST_ContextVK::initializeDrawingContext()
vkGetDeviceQueue(m_device, m_queue_family_graphic, 0, &m_graphic_queue);
createCommandPools();
if (use_window_surface) {
vkGetDeviceQueue(m_device, m_queue_family_present, 0, &m_present_queue);
createSwapchain();
}
else {
createGraphicsCommandBuffer();
}
return GHOST_kSuccess;
}

17
intern/ghost/intern/GHOST_ContextVK.h
View File

@ -113,18 +113,17 @@ class GHOST_ContextVK : public GHOST_Context {
GHOST_TSuccess getVulkanHandles(void *r_instance,
void *r_physical_device,
void *r_device,
uint32_t *r_graphic_queue_family);
uint32_t *r_graphic_queue_family,
void *r_queue);
GHOST_TSuccess getVulkanCommandBuffer(void *r_command_buffer);
/**
* Gets the Vulkan framebuffer related resource handles associated with the Vulkan context.
* Needs to be called after each swap events as the framebuffer will change.
* \return A boolean success indicator.
*/
GHOST_TSuccess getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id);
GHOST_TSuccess getVulkanBackbuffer(
void *image, void *framebuffer, void *render_pass, void *extent, uint32_t *fb_id);
/**
* Sets the swap interval for swapBuffers.
@ -200,6 +199,8 @@ class GHOST_ContextVK : public GHOST_Context {
GHOST_TSuccess pickPhysicalDevice(std::vector<const char *> required_exts);
GHOST_TSuccess createSwapchain();
GHOST_TSuccess destroySwapchain();
GHOST_TSuccess createCommandBuffers();
GHOST_TSuccess createCommandPools();
GHOST_TSuccess createGraphicsCommandBuffers();
GHOST_TSuccess createGraphicsCommandBuffer();
GHOST_TSuccess recordCommandBuffers();
};

3
intern/ghost/intern/GHOST_Window.cpp
View File

@ -109,13 +109,12 @@ uint GHOST_Window::getDefaultFramebuffer()
GHOST_TSuccess GHOST_Window::getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id)
{
return m_context->getVulkanBackbuffer(
image, framebuffer, command_buffer, render_pass, extent, fb_id);
image, framebuffer, render_pass, extent, fb_id);
}
GHOST_TSuccess GHOST_Window::activateDrawingContext()

8
intern/ghost/intern/GHOST_Window.h
View File

@ -274,12 +274,8 @@ class GHOST_Window : public GHOST_IWindow {
* Needs to be called after each swap events as the framebuffer will change.
* \return A boolean success indicator.
*/
virtual GHOST_TSuccess getVulkanBackbuffer(void *image,
void *framebuffer,
void *command_buffer,
void *render_pass,
void *extent,
uint32_t *fb_id) override;
virtual GHOST_TSuccess getVulkanBackbuffer(
void *image, void *framebuffer, void *render_pass, void *extent, uint32_t *fb_id) override;
/**
* Returns the window user data.

18
source/blender/gpu/CMakeLists.txt
View File

@ -191,16 +191,24 @@ set(OPENGL_SRC
set(VULKAN_SRC
vulkan/vk_backend.cc
vulkan/vk_batch.cc
vulkan/vk_buffer.cc
vulkan/vk_context.cc
vulkan/vk_command_buffer.cc
vulkan/vk_common.cc
vulkan/vk_descriptor_pools.cc
vulkan/vk_descriptor_set.cc
vulkan/vk_drawlist.cc
vulkan/vk_fence.cc
vulkan/vk_framebuffer.cc
vulkan/vk_index_buffer.cc
vulkan/vk_pipeline.cc
vulkan/vk_memory.cc
vulkan/vk_pixel_buffer.cc
vulkan/vk_query.cc
vulkan/vk_shader.cc
vulkan/vk_shader_interface.cc
vulkan/vk_shader_log.cc
vulkan/vk_state_manager.cc
vulkan/vk_storage_buffer.cc
vulkan/vk_texture.cc
vulkan/vk_uniform_buffer.cc
@ -208,16 +216,24 @@ set(VULKAN_SRC
vulkan/vk_backend.hh
vulkan/vk_batch.hh
vulkan/vk_buffer.hh
vulkan/vk_context.hh
vulkan/vk_command_buffer.hh
vulkan/vk_common.hh
vulkan/vk_descriptor_pools.hh
vulkan/vk_descriptor_set.hh
vulkan/vk_drawlist.hh
vulkan/vk_fence.hh
vulkan/vk_framebuffer.hh
vulkan/vk_index_buffer.hh
vulkan/vk_pipeline.hh
vulkan/vk_memory.hh
vulkan/vk_pixel_buffer.hh
vulkan/vk_query.hh
vulkan/vk_shader.hh
vulkan/vk_shader_interface.hh
vulkan/vk_shader_log.hh
vulkan/vk_state_manager.hh
vulkan/vk_storage_buffer.hh
vulkan/vk_texture.hh
vulkan/vk_uniform_buffer.hh
@ -510,6 +526,7 @@ set(GLSL_SRC_TEST
tests/shaders/gpu_compute_1d_test.glsl
tests/shaders/gpu_compute_2d_test.glsl
tests/shaders/gpu_compute_ibo_test.glsl
tests/shaders/gpu_compute_ssbo_test.glsl
tests/shaders/gpu_compute_vbo_test.glsl
tests/shaders/gpu_compute_dummy_test.glsl
)
@ -787,6 +804,7 @@ if(WITH_GTESTS)
tests/gpu_index_buffer_test.cc
tests/gpu_shader_builtin_test.cc
tests/gpu_shader_test.cc
tests/gpu_storage_buffer_test.cc
tests/gpu_testing.hh
)

8
source/blender/gpu/intern/gpu_shader_interface.hh
View File

@ -26,6 +26,14 @@ namespace blender::gpu {
typedef struct ShaderInput {
uint32_t name_offset;
uint32_t name_hash;
/**
* Location is openGl legacy and its legacy usages should be phased out in Blender 3.7.
*
* Vulkan backend use location to encode the descriptor set binding. This binding is different
* than the binding stored in the binding attribute. In Vulkan the binding inside a descriptor
* set must be unique. In future the location will also be used to select the right descriptor
* set.
*/
int32_t location;
/** Defined at interface creation or in shader. Only for Samplers, UBOs and Vertex Attributes. */
int32_t binding;

2
source/blender/gpu/intern/gpu_storage_buffer_private.hh
View File

@ -14,6 +14,8 @@ struct GPUStorageBuf;
namespace blender {
namespace gpu {
class VertBuf;
#ifdef DEBUG
# define DEBUG_NAME_LEN 64
#else

6
source/blender/gpu/shaders/infos/gpu_shader_test_info.hh
View File

@ -42,6 +42,12 @@ GPU_SHADER_CREATE_INFO(gpu_compute_vbo_test)
.compute_source("gpu_compute_vbo_test.glsl")
.do_static_compilation(true);
GPU_SHADER_CREATE_INFO(gpu_compute_ssbo_test)
.local_group_size(1)
.storage_buf(0, Qualifier::WRITE, "int", "data_out[]")
.compute_source("gpu_compute_ssbo_test.glsl")
.do_static_compilation(true);
GPU_SHADER_CREATE_INFO(gpu_compute_ssbo_binding_test)
.local_group_size(1)
.storage_buf(0, Qualifier::WRITE, "int", "data0[]")

8
source/blender/gpu/tests/gpu_shader_test.cc
View File

@ -212,7 +212,7 @@ GPU_TEST(gpu_shader_compute_ibo)
static void test_gpu_shader_compute_ssbo()
{
if (!GPU_compute_shader_support()) {
if (!GPU_compute_shader_support() && !GPU_shader_storage_buffer_objects_support()) {
/* We can't test as a the platform does not support compute shaders. */
std::cout << "Skipping compute shader test: platform not supported";
return;
@ -221,14 +221,14 @@ static void test_gpu_shader_compute_ssbo()
static constexpr uint SIZE = 128;
/* Build compute shader. */
GPUShader *shader = GPU_shader_create_from_info_name("gpu_compute_ibo_test");
GPUShader *shader = GPU_shader_create_from_info_name("gpu_compute_ssbo_test");
EXPECT_NE(shader, nullptr);
GPU_shader_bind(shader);
/* Construct IBO. */
GPUStorageBuf *ssbo = GPU_storagebuf_create_ex(
SIZE * sizeof(uint32_t), nullptr, GPU_USAGE_DEVICE_ONLY, __func__);
GPU_storagebuf_bind(ssbo, GPU_shader_get_ssbo_binding(shader, "out_indices"));
GPU_storagebuf_bind(ssbo, GPU_shader_get_ssbo_binding(shader, "data_out"));
/* Dispatch compute task. */
GPU_compute_dispatch(shader, SIZE, 1, 1);
@ -240,7 +240,7 @@ static void test_gpu_shader_compute_ssbo()
uint32_t data[SIZE];
GPU_storagebuf_read(ssbo, data);
for (int index = 0; index < SIZE; index++) {
uint32_t expected = index;
uint32_t expected = index * 4;
EXPECT_EQ(data[index], expected);
}

50
source/blender/gpu/tests/gpu_storage_buffer_test.cc Normal file
View File

@ -0,0 +1,50 @@
/* SPDX-License-Identifier: Apache-2.0 */
#include "testing/testing.h"
#include "GPU_storage_buffer.h"
#include "BLI_vector.hh"
#include "gpu_testing.hh"
namespace blender::gpu::tests {
constexpr size_t SIZE = 128;
constexpr size_t SIZE_IN_BYTES = SIZE * sizeof(int);
static Vector<int32_t> test_data()
{
Vector<int32_t> data;
for (int i : IndexRange(SIZE)) {
data.append(i);
}
return data;
}
static void test_gpu_storage_buffer_create_update_read()
{
GPUStorageBuf *ssbo = GPU_storagebuf_create_ex(
SIZE_IN_BYTES, nullptr, GPU_USAGE_STATIC, __func__);
EXPECT_NE(ssbo, nullptr);
/* Upload some dummy data. */
const Vector<int32_t> data = test_data();
GPU_storagebuf_update(ssbo, data.data());
/* Read back data from SSBO. */
Vector<int32_t> read_data;
read_data.resize(SIZE, 0);
GPU_storagebuf_read(ssbo, read_data.data());
/* Check if data is the same.*/
for (int i : IndexRange(SIZE)) {
EXPECT_EQ(data[i], read_data[i]);
}
GPU_storagebuf_free(ssbo);
}
GPU_TEST(gpu_storage_buffer_create_update_read);
} // namespace blender::gpu::tests

5
source/blender/gpu/tests/gpu_testing.cc
View File

@ -15,18 +15,21 @@ namespace blender::gpu {
void GPUTest::SetUp()
{
GPU_backend_type_selection_set(gpu_backend_type);
GHOST_GLSettings glSettings = {0};
GHOST_GLSettings glSettings = {};
glSettings.context_type = draw_context_type;
glSettings.flags = GHOST_glDebugContext;
CLG_init();
ghost_system = GHOST_CreateSystem();
ghost_context = GHOST_CreateOpenGLContext(ghost_system, glSettings);
GHOST_ActivateOpenGLContext(ghost_context);
context = GPU_context_create(nullptr, ghost_context);
GPU_init();
GPU_context_begin_frame(context);
}
void GPUTest::TearDown()
{
GPU_context_end_frame(context);
GPU_exit();
GPU_context_discard(context);
GHOST_DisposeOpenGLContext(ghost_system, ghost_context);

5
source/blender/gpu/tests/shaders/gpu_compute_ssbo_test.glsl Normal file
View File

@ -0,0 +1,5 @@
void main()
{
int store_index = int(gl_GlobalInvocationID.x);
data_out[store_index] = store_index * 4;
}

15
source/blender/gpu/vulkan/vk_backend.cc
View File

@ -60,8 +60,17 @@ void VKBackend::samplers_update()
{
}
void VKBackend::compute_dispatch(int /*groups_x_len*/, int /*groups_y_len*/, int /*groups_z_len*/)
void VKBackend::compute_dispatch(int groups_x_len, int groups_y_len, int groups_z_len)
{
VKContext &context = *VKContext::get();
VKShader *shader = static_cast<VKShader *>(context.shader);
VKCommandBuffer &command_buffer = context.command_buffer_get();
VKPipeline &pipeline = shader->pipeline_get();
VKDescriptorSet &descriptor_set = pipeline.descriptor_set_get();
descriptor_set.update(context.device_get());
command_buffer.bind(
descriptor_set, shader->vk_pipeline_layout_get(), VK_PIPELINE_BIND_POINT_COMPUTE);
command_buffer.dispatch(groups_x_len, groups_y_len, groups_z_len);
}
void VKBackend::compute_dispatch_indirect(StorageBuf * /*indirect_buf*/)
@ -123,9 +132,9 @@ UniformBuf *VKBackend::uniformbuf_alloc(int size, const char *name)
return new VKUniformBuffer(size, name);
}
StorageBuf *VKBackend::storagebuf_alloc(int size, GPUUsageType /*usage*/, const char *name)
StorageBuf *VKBackend::storagebuf_alloc(int size, GPUUsageType usage, const char *name)
{
return new VKStorageBuffer(size, name);
return new VKStorageBuffer(size, usage, name);
}
VertBuf *VKBackend::vertbuf_alloc()

7
source/blender/gpu/vulkan/vk_backend.hh
View File

@ -9,11 +9,8 @@
#include "gpu_backend.hh"
#ifdef __APPLE__
# include <MoltenVK/vk_mvk_moltenvk.h>
#else
# include <vulkan/vulkan.h>
#endif
#include "vk_common.hh"
#include "shaderc/shaderc.hpp"
namespace blender::gpu {

107
source/blender/gpu/vulkan/vk_buffer.cc Normal file
View File

@ -0,0 +1,107 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_buffer.hh"
namespace blender::gpu {
VKBuffer::~VKBuffer()
{
VKContext &context = *VKContext::get();
free(context);
}
bool VKBuffer::is_allocated() const
{
return allocation_ != VK_NULL_HANDLE;
}
static VmaAllocationCreateFlagBits vma_allocation_flags(GPUUsageType usage)
{
switch (usage) {
case GPU_USAGE_STATIC:
return static_cast<VmaAllocationCreateFlagBits>(
VMA_ALLOCATION_CREATE_HOST_ACCESS_RANDOM_BIT | VMA_ALLOCATION_CREATE_MAPPED_BIT);
case GPU_USAGE_DYNAMIC:
return static_cast<VmaAllocationCreateFlagBits>(
VMA_ALLOCATION_CREATE_HOST_ACCESS_RANDOM_BIT | VMA_ALLOCATION_CREATE_MAPPED_BIT);
case GPU_USAGE_DEVICE_ONLY:
return static_cast<VmaAllocationCreateFlagBits>(
VMA_ALLOCATION_CREATE_HOST_ACCESS_RANDOM_BIT |
VMA_ALLOCATION_CREATE_DEDICATED_MEMORY_BIT);
case GPU_USAGE_FLAG_BUFFER_TEXTURE_ONLY:
case GPU_USAGE_STREAM:
break;
}
BLI_assert_msg(false, "Unimplemented GPUUsageType");
return static_cast<VmaAllocationCreateFlagBits>(VMA_ALLOCATION_CREATE_HOST_ACCESS_RANDOM_BIT |
VMA_ALLOCATION_CREATE_MAPPED_BIT);
}
bool VKBuffer::create(VKContext &context,
int64_t size_in_bytes,
GPUUsageType usage,
VkBufferUsageFlagBits buffer_usage)
{
BLI_assert(!is_allocated());
size_in_bytes_ = size_in_bytes;
VmaAllocator allocator = context.mem_allocator_get();
VkBufferCreateInfo create_info = {};
create_info.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
create_info.flags = 0;
create_info.size = size_in_bytes;
create_info.usage = buffer_usage;
/* We use the same command queue for the compute and graphics pipeline, so it is safe to use
* exclusive resource handling. */
create_info.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
create_info.queueFamilyIndexCount = 1;
create_info.pQueueFamilyIndices = context.queue_family_ptr_get();
VmaAllocationCreateInfo vma_create_info = {};
vma_create_info.flags = vma_allocation_flags(usage);
vma_create_info.priority = 1.0f;
vma_create_info.usage = VMA_MEMORY_USAGE_AUTO;
VkResult result = vmaCreateBuffer(
allocator, &create_info, &vma_create_info, &vk_buffer_, &allocation_, nullptr);
return result == VK_SUCCESS;
}
bool VKBuffer::update(VKContext &context, const void *data)
{
void *mapped_memory;
bool result = map(context, &mapped_memory);
if (result) {
memcpy(mapped_memory, data, size_in_bytes_);
unmap(context);
}
return result;
}
bool VKBuffer::map(VKContext &context, void **r_mapped_memory) const
{
VmaAllocator allocator = context.mem_allocator_get();
VkResult result = vmaMapMemory(allocator, allocation_, r_mapped_memory);
return result == VK_SUCCESS;
}
void VKBuffer::unmap(VKContext &context) const
{
VmaAllocator allocator = context.mem_allocator_get();
vmaUnmapMemory(allocator, allocation_);
}
bool VKBuffer::free(VKContext &context)
{
VmaAllocator allocator = context.mem_allocator_get();
vmaDestroyBuffer(allocator, vk_buffer_, allocation_);
return true;
}
} // namespace blender::gpu

53
source/blender/gpu/vulkan/vk_buffer.hh Normal file
View File

@ -0,0 +1,53 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "gpu_context_private.hh"
#include "vk_common.hh"
#include "vk_context.hh"
#include "vk_mem_alloc.h"
namespace blender::gpu {
/**
* Class for handing vulkan buffers (allocation/updating/binding).
*/
class VKBuffer {
int64_t size_in_bytes_;
VkBuffer vk_buffer_ = VK_NULL_HANDLE;
VmaAllocation allocation_ = VK_NULL_HANDLE;
public:
VKBuffer() = default;
virtual ~VKBuffer();
/** Has this buffer been allocated? */
bool is_allocated() const;
bool create(VKContext &context,
int64_t size,
GPUUsageType usage,
VkBufferUsageFlagBits buffer_usage);
bool update(VKContext &context, const void *data);
bool free(VKContext &context);
bool map(VKContext &context, void **r_mapped_memory) const;
void unmap(VKContext &context) const;
int64_t size_in_bytes() const
{
return size_in_bytes_;
}
VkBuffer vk_handle() const
{
return vk_buffer_;
}
};
} // namespace blender::gpu

144
source/blender/gpu/vulkan/vk_command_buffer.cc Normal file
View File

@ -0,0 +1,144 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_command_buffer.hh"
#include "vk_buffer.hh"
#include "vk_context.hh"
#include "vk_memory.hh"
#include "vk_texture.hh"
#include "BLI_assert.h"
namespace blender::gpu {
VKCommandBuffer::~VKCommandBuffer()
{
if (vk_device_ != VK_NULL_HANDLE) {
VK_ALLOCATION_CALLBACKS;
vkDestroyFence(vk_device_, vk_fence_, vk_allocation_callbacks);
vk_fence_ = VK_NULL_HANDLE;
}
}
void VKCommandBuffer::init(const VkDevice vk_device,
const VkQueue vk_queue,
VkCommandBuffer vk_command_buffer)
{
vk_device_ = vk_device;
vk_queue_ = vk_queue;
vk_command_buffer_ = vk_command_buffer;
if (vk_fence_ == VK_NULL_HANDLE) {
VK_ALLOCATION_CALLBACKS;
VkFenceCreateInfo fenceInfo{};
fenceInfo.sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO;
fenceInfo.flags = VK_FENCE_CREATE_SIGNALED_BIT;
vkCreateFence(vk_device_, &fenceInfo, vk_allocation_callbacks, &vk_fence_);
}
}
void VKCommandBuffer::begin_recording()
{
vkWaitForFences(vk_device_, 1, &vk_fence_, VK_TRUE, UINT64_MAX);
vkResetFences(vk_device_, 1, &vk_fence_);
vkResetCommandBuffer(vk_command_buffer_, 0);
VkCommandBufferBeginInfo begin_info = {};
begin_info.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO;
vkBeginCommandBuffer(vk_command_buffer_, &begin_info);
}
void VKCommandBuffer::end_recording()
{
vkEndCommandBuffer(vk_command_buffer_);
}
void VKCommandBuffer::bind(const VKPipeline &pipeline, VkPipelineBindPoint bind_point)
{
vkCmdBindPipeline(vk_command_buffer_, bind_point, pipeline.vk_handle());
}
void VKCommandBuffer::bind(const VKDescriptorSet &descriptor_set,
const VkPipelineLayout vk_pipeline_layout,
VkPipelineBindPoint bind_point)
{
VkDescriptorSet vk_descriptor_set = descriptor_set.vk_handle();
vkCmdBindDescriptorSets(
vk_command_buffer_, bind_point, vk_pipeline_layout, 0, 1, &vk_descriptor_set, 0, 0);
}
void VKCommandBuffer::copy(VKBuffer &dst_buffer,
VKTexture &src_texture,
Span<VkBufferImageCopy> regions)
{
vkCmdCopyImageToBuffer(vk_command_buffer_,
src_texture.vk_image_handle(),
VK_IMAGE_LAYOUT_GENERAL,
dst_buffer.vk_handle(),
regions.size(),
regions.data());
}
void VKCommandBuffer::pipeline_barrier(VkPipelineStageFlags source_stages,
VkPipelineStageFlags destination_stages)
{
vkCmdPipelineBarrier(vk_command_buffer_,
source_stages,
destination_stages,
0,
0,
nullptr,
0,
nullptr,
0,
nullptr);
}
void VKCommandBuffer::pipeline_barrier(Span<VkImageMemoryBarrier> image_memory_barriers)
{
vkCmdPipelineBarrier(vk_command_buffer_,
VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT,
VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT,
VK_DEPENDENCY_BY_REGION_BIT,
0,
nullptr,
0,
nullptr,
image_memory_barriers.size(),
image_memory_barriers.data());
}
void VKCommandBuffer::dispatch(int groups_x_len, int groups_y_len, int groups_z_len)
{
vkCmdDispatch(vk_command_buffer_, groups_x_len, groups_y_len, groups_z_len);
}
void VKCommandBuffer::submit()
{
end_recording();
encode_recorded_commands();
submit_encoded_commands();
begin_recording();
}
void VKCommandBuffer::encode_recorded_commands()
{
/* Intentionally not implemented. For the graphics pipeline we want to extract the
* resources and its usages so we can encode multiple commands in the same command buffer with
* the correct synchorinzations. */
}
void VKCommandBuffer::submit_encoded_commands()
{
VkSubmitInfo submit_info = {};
submit_info.sType = VK_STRUCTURE_TYPE_SUBMIT_INFO;
submit_info.commandBufferCount = 1;
submit_info.pCommandBuffers = &vk_command_buffer_;
vkQueueSubmit(vk_queue_, 1, &submit_info, vk_fence_);
}
} // namespace blender::gpu

54
source/blender/gpu/vulkan/vk_command_buffer.hh Normal file
View File

@ -0,0 +1,54 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "vk_common.hh"
#include "vk_pipeline.hh"
namespace blender::gpu {
class VKBuffer;
class VKTexture;
/** Command buffer to keep track of the life-time of a command buffer.*/
class VKCommandBuffer : NonCopyable, NonMovable {
/** None owning handle to the command buffer and device. Handle is owned by `GHOST_ContextVK`.*/
VkDevice vk_device_ = VK_NULL_HANDLE;
VkCommandBuffer vk_command_buffer_ = VK_NULL_HANDLE;
VkQueue vk_queue_ = VK_NULL_HANDLE;
/** Owning handles */
VkFence vk_fence_ = VK_NULL_HANDLE;
public:
virtual ~VKCommandBuffer();
void init(const VkDevice vk_device, const VkQueue vk_queue, VkCommandBuffer vk_command_buffer);
void begin_recording();
void end_recording();
void bind(const VKPipeline &vk_pipeline, VkPipelineBindPoint bind_point);
void bind(const VKDescriptorSet &descriptor_set,
const VkPipelineLayout vk_pipeline_layout,
VkPipelineBindPoint bind_point);
void dispatch(int groups_x_len, int groups_y_len, int groups_z_len);
/* Copy the contents of a texture mip level to the dst buffer.*/
void copy(VKBuffer &dst_buffer, VKTexture &src_texture, Span<VkBufferImageCopy> regions);
void pipeline_barrier(VkPipelineStageFlags source_stages,
VkPipelineStageFlags destination_stages);
void pipeline_barrier(Span<VkImageMemoryBarrier> image_memory_barriers);
/**
* 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();
private:
void encode_recorded_commands();
void submit_encoded_commands();
};
} // namespace blender::gpu

197
source/blender/gpu/vulkan/vk_common.cc Normal file
View File

@ -0,0 +1,197 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_common.hh"
namespace blender::gpu {
VkImageAspectFlagBits to_vk_image_aspect_flag_bits(const eGPUTextureFormat format)
{
switch (format) {
case GPU_RGBA32F:
case GPU_RGBA8UI:
case GPU_RGBA8I:
case GPU_RGBA8:
case GPU_RGBA32UI:
case GPU_RGBA32I:
case GPU_RGBA16UI:
case GPU_RGBA16I:
case GPU_RGBA16F:
case GPU_RGBA16:
case GPU_RG8UI:
case GPU_RG8I:
case GPU_RG8:
case GPU_RG32UI:
case GPU_RG32I:
case GPU_RG32F:
case GPU_RG16UI:
case GPU_RG16I:
case GPU_RG16F:
case GPU_RG16:
case GPU_R8UI:
case GPU_R8I:
case GPU_R8:
case GPU_R32UI:
case GPU_R32I:
case GPU_R32F:
case GPU_R16UI:
case GPU_R16I:
case GPU_R16F:
case GPU_R16:
case GPU_RGB10_A2:
case GPU_R11F_G11F_B10F:
case GPU_SRGB8_A8:
case GPU_RGB16F:
case GPU_SRGB8_A8_DXT1:
case GPU_SRGB8_A8_DXT3:
case GPU_SRGB8_A8_DXT5:
case GPU_RGBA8_DXT1:
case GPU_RGBA8_DXT3:
case GPU_RGBA8_DXT5:
return VK_IMAGE_ASPECT_COLOR_BIT;
case GPU_DEPTH32F_STENCIL8:
case GPU_DEPTH24_STENCIL8:
return static_cast<VkImageAspectFlagBits>(VK_IMAGE_ASPECT_DEPTH_BIT |
VK_IMAGE_ASPECT_STENCIL_BIT);
case GPU_DEPTH_COMPONENT24:
case GPU_DEPTH_COMPONENT16:
return VK_IMAGE_ASPECT_DEPTH_BIT;
case GPU_DEPTH_COMPONENT32F:
/* Not supported by Vulkan*/
BLI_assert_unreachable();
}
return static_cast<VkImageAspectFlagBits>(0);
}
VkFormat to_vk_format(const eGPUTextureFormat format)
{
switch (format) {
case GPU_RGBA32F:
return VK_FORMAT_R32G32B32A32_SFLOAT;
case GPU_RGBA8UI:
case GPU_RGBA8I:
case GPU_RGBA8:
case GPU_RGBA32UI:
case GPU_RGBA32I:
case GPU_RGBA16UI:
case GPU_RGBA16I:
case GPU_RGBA16F:
case GPU_RGBA16:
case GPU_RG8UI:
case GPU_RG8I:
case GPU_RG8:
case GPU_RG32UI:
case GPU_RG32I:
case GPU_RG32F:
case GPU_RG16UI:
case GPU_RG16I:
case GPU_RG16F:
case GPU_RG16:
case GPU_R8UI:
case GPU_R8I:
case GPU_R8:
case GPU_R32UI:
case GPU_R32I:
case GPU_R32F:
case GPU_R16UI:
case GPU_R16I:
case GPU_R16F:
case GPU_R16:
case GPU_RGB10_A2:
case GPU_R11F_G11F_B10F:
case GPU_DEPTH32F_STENCIL8:
case GPU_DEPTH24_STENCIL8:
case GPU_SRGB8_A8:
/* Texture only format */
case GPU_RGB16F:
/* Special formats texture only */
case GPU_SRGB8_A8_DXT1:
case GPU_SRGB8_A8_DXT3:
case GPU_SRGB8_A8_DXT5:
case GPU_RGBA8_DXT1:
case GPU_RGBA8_DXT3:
case GPU_RGBA8_DXT5:
/* Depth Formats */
case GPU_DEPTH_COMPONENT32F:
case GPU_DEPTH_COMPONENT24:
case GPU_DEPTH_COMPONENT16:
BLI_assert_unreachable();
}
return VK_FORMAT_UNDEFINED;
}
VkImageType to_vk_image_type(const eGPUTextureType type)
{
switch (type) {
case GPU_TEXTURE_1D:
case GPU_TEXTURE_BUFFER:
case GPU_TEXTURE_1D_ARRAY:
return VK_IMAGE_TYPE_1D;
case GPU_TEXTURE_2D:
case GPU_TEXTURE_2D_ARRAY:
return VK_IMAGE_TYPE_2D;
case GPU_TEXTURE_3D:
case GPU_TEXTURE_CUBE:
case GPU_TEXTURE_CUBE_ARRAY:
return VK_IMAGE_TYPE_3D;
case GPU_TEXTURE_ARRAY:
/* GPU_TEXTURE_ARRAY should always be used together with 1D, 2D, or CUBE*/
BLI_assert_unreachable();
break;
}
return VK_IMAGE_TYPE_1D;
}
VkImageViewType to_vk_image_view_type(const eGPUTextureType type)
{
switch (type) {
case GPU_TEXTURE_1D:
case GPU_TEXTURE_BUFFER:
return VK_IMAGE_VIEW_TYPE_1D;
case GPU_TEXTURE_2D:
return VK_IMAGE_VIEW_TYPE_2D;
case GPU_TEXTURE_3D:
return VK_IMAGE_VIEW_TYPE_3D;
case GPU_TEXTURE_CUBE:
return VK_IMAGE_VIEW_TYPE_CUBE;
case GPU_TEXTURE_1D_ARRAY:
return VK_IMAGE_VIEW_TYPE_1D_ARRAY;
case GPU_TEXTURE_2D_ARRAY:
return VK_IMAGE_VIEW_TYPE_2D_ARRAY;
case GPU_TEXTURE_CUBE_ARRAY:
return VK_IMAGE_VIEW_TYPE_CUBE_ARRAY;
case GPU_TEXTURE_ARRAY:
/* GPU_TEXTURE_ARRAY should always be used together with 1D, 2D, or CUBE*/
BLI_assert_unreachable();
break;
}
return VK_IMAGE_VIEW_TYPE_1D;
}
VkComponentMapping to_vk_component_mapping(const eGPUTextureFormat /*format*/)
{
/* TODO: this should map to OpenGL defaults based on the eGPUTextureFormat. The implementation of
* this function will be implemented when implementing other parts of VKTexture.*/
VkComponentMapping component_mapping;
component_mapping.r = VK_COMPONENT_SWIZZLE_R;
component_mapping.g = VK_COMPONENT_SWIZZLE_G;
component_mapping.b = VK_COMPONENT_SWIZZLE_B;
component_mapping.a = VK_COMPONENT_SWIZZLE_A;
return component_mapping;
}
} // namespace blender::gpu

26
source/blender/gpu/vulkan/vk_common.hh Normal file
View File

@ -0,0 +1,26 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#ifdef __APPLE__
# include <MoltenVK/vk_mvk_moltenvk.h>
#else
# include <vulkan/vulkan.h>
#endif
#include "gpu_texture_private.hh"
namespace blender::gpu {
VkImageAspectFlagBits to_vk_image_aspect_flag_bits(const eGPUTextureFormat format);
VkFormat to_vk_format(const eGPUTextureFormat format);
VkComponentMapping to_vk_component_mapping(const eGPUTextureFormat format);
VkImageViewType to_vk_image_view_type(const eGPUTextureType type);
VkImageType to_vk_image_type(const eGPUTextureType type);
} // namespace blender::gpu

38
source/blender/gpu/vulkan/vk_context.cc
View File

@ -8,6 +8,8 @@
#include "vk_context.hh"
#include "vk_backend.hh"
#include "vk_memory.hh"
#include "vk_state_manager.hh"
#include "GHOST_C-api.h"
@ -15,25 +17,34 @@ namespace blender::gpu {
VKContext::VKContext(void *ghost_window, void *ghost_context)
{
VK_ALLOCATION_CALLBACKS;
ghost_window_ = ghost_window;
if (ghost_window) {
ghost_context = GHOST_GetDrawingContext((GHOST_WindowHandle)ghost_window);
}
ghost_context_ = ghost_context;
GHOST_GetVulkanHandles((GHOST_ContextHandle)ghost_context,
&instance_,
&physical_device_,
&device_,
&graphic_queue_family_);
&vk_instance_,
&vk_physical_device_,
&vk_device_,
&vk_queue_family_,
&vk_queue_);
/* Initialize the memory allocator. */
VmaAllocatorCreateInfo info = {};
/* Should use same vulkan version as GHOST. */
info.vulkanApiVersion = VK_API_VERSION_1_2;
info.physicalDevice = physical_device_;
info.device = device_;
info.instance = instance_;
/* Should use same vulkan version as GHOST (1.2), but set to 1.0 as 1.2 requires
* correct extensions and functions to be found by VMA, which isn't working as expected and
* requires more research. To continue development we lower the API to version 1.0.*/
info.vulkanApiVersion = VK_API_VERSION_1_0;
info.physicalDevice = vk_physical_device_;
info.device = vk_device_;
info.instance = vk_instance_;
info.pAllocationCallbacks = vk_allocation_callbacks;
vmaCreateAllocator(&info, &mem_allocator_);
descriptor_pools_.init(vk_device_);
state_manager = new VKStateManager();
VKBackend::capabilities_init(*this);
}
@ -53,18 +64,27 @@ void VKContext::deactivate()
void VKContext::begin_frame()
{
VkCommandBuffer command_buffer = VK_NULL_HANDLE;
GHOST_GetVulkanCommandBuffer(static_cast<GHOST_ContextHandle>(ghost_context_), &command_buffer);
command_buffer_.init(vk_device_, vk_queue_, command_buffer);
command_buffer_.begin_recording();
descriptor_pools_.reset();
}
void VKContext::end_frame()
{
command_buffer_.end_recording();
}
void VKContext::flush()
{
command_buffer_.submit();
}
void VKContext::finish()
{
command_buffer_.submit();
}
void VKContext::memory_statistics_get(int * /*total_mem*/, int * /*free_mem*/)

49
source/blender/gpu/vulkan/vk_context.hh
View File

@ -9,26 +9,28 @@
#include "gpu_context_private.hh"
#include "vk_mem_alloc.h"
#include "vk_command_buffer.hh"
#include "vk_descriptor_pools.hh"
#ifdef __APPLE__
# include <MoltenVK/vk_mvk_moltenvk.h>
#else
# include <vulkan/vulkan.h>
#endif
#include "vk_mem_alloc.h"
namespace blender::gpu {
class VKContext : public Context {
private:
/** Copies of the handles owned by the GHOST context. */
VkInstance instance_ = VK_NULL_HANDLE;
VkPhysicalDevice physical_device_ = VK_NULL_HANDLE;
VkDevice device_ = VK_NULL_HANDLE;
uint32_t graphic_queue_family_ = 0;
VkInstance vk_instance_ = VK_NULL_HANDLE;
VkPhysicalDevice vk_physical_device_ = VK_NULL_HANDLE;
VkDevice vk_device_ = VK_NULL_HANDLE;
VKCommandBuffer command_buffer_;
uint32_t vk_queue_family_ = 0;
VkQueue vk_queue_ = VK_NULL_HANDLE;
/** Allocator used for texture and buffers and other resources. */
VmaAllocator mem_allocator_ = VK_NULL_HANDLE;
VKDescriptorPools descriptor_pools_;
void *ghost_context_;
public:
VKContext(void *ghost_window, void *ghost_context);
@ -52,9 +54,34 @@ class VKContext : public Context {
return static_cast<VKContext *>(Context::get());
}
VkPhysicalDevice physical_device_get() const
{
return vk_physical_device_;
}
VkDevice device_get() const
{
return device_;
return vk_device_;
}
VKCommandBuffer &command_buffer_get()
{
return command_buffer_;
}
VkQueue queue_get() const
{
return vk_queue_;
}
const uint32_t *queue_family_ptr_get() const
{
return &vk_queue_family_;
}
VKDescriptorPools &descriptor_pools_get()
{
return descriptor_pools_;
}
VmaAllocator mem_allocator_get() const

116
source/blender/gpu/vulkan/vk_descriptor_pools.cc Normal file
View File

@ -0,0 +1,116 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_descriptor_pools.hh"
#include "vk_memory.hh"
namespace blender::gpu {
VKDescriptorPools::VKDescriptorPools()
{
}
VKDescriptorPools::~VKDescriptorPools()
{
VK_ALLOCATION_CALLBACKS
for (const VkDescriptorPool vk_descriptor_pool : pools_) {
BLI_assert(vk_device_ != VK_NULL_HANDLE);
vkDestroyDescriptorPool(vk_device_, vk_descriptor_pool, vk_allocation_callbacks);
}
vk_device_ = VK_NULL_HANDLE;
}
void VKDescriptorPools::init(const VkDevice vk_device)
{
BLI_assert(vk_device_ == VK_NULL_HANDLE);
vk_device_ = vk_device;
add_new_pool();
}
void VKDescriptorPools::reset()
{
active_pool_index_ = 0;
}
void VKDescriptorPools::add_new_pool()
{
VK_ALLOCATION_CALLBACKS
Vector<VkDescriptorPoolSize> pool_sizes = {
{VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, POOL_SIZE_STORAGE_BUFFER},
{VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, POOL_SIZE_STORAGE_IMAGE},
{VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, POOL_SIZE_COMBINED_IMAGE_SAMPLER},
{VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, POOL_SIZE_UNIFORM_BUFFER},
};
VkDescriptorPoolCreateInfo pool_info = {};
pool_info.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO;
pool_info.flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT;
pool_info.maxSets = POOL_SIZE_DESCRIPTOR_SETS;
pool_info.poolSizeCount = pool_sizes.size();
pool_info.pPoolSizes = pool_sizes.data();
VkDescriptorPool descriptor_pool = VK_NULL_HANDLE;
VkResult result = vkCreateDescriptorPool(
vk_device_, &pool_info, vk_allocation_callbacks, &descriptor_pool);
UNUSED_VARS(result);
pools_.append(descriptor_pool);
}
VkDescriptorPool VKDescriptorPools::active_pool_get()
{
BLI_assert(!pools_.is_empty());
return pools_[active_pool_index_];
}
void VKDescriptorPools::activate_next_pool()
{
BLI_assert(!is_last_pool_active());
active_pool_index_ += 1;
}
void VKDescriptorPools::activate_last_pool()
{
active_pool_index_ = pools_.size() - 1;
}
bool VKDescriptorPools::is_last_pool_active()
{
return active_pool_index_ == pools_.size() - 1;
}
VKDescriptorSet VKDescriptorPools::allocate(const VkDescriptorSetLayout &descriptor_set_layout)
{
VkDescriptorSetAllocateInfo allocate_info = {};
VkDescriptorPool pool = active_pool_get();
allocate_info.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO;
allocate_info.descriptorPool = pool;
allocate_info.descriptorSetCount = 1;
allocate_info.pSetLayouts = &descriptor_set_layout;
VkDescriptorSet vk_descriptor_set = VK_NULL_HANDLE;
VkResult result = vkAllocateDescriptorSets(vk_device_, &allocate_info, &vk_descriptor_set);
if (result == VK_ERROR_OUT_OF_POOL_MEMORY) {
if (is_last_pool_active()) {
add_new_pool();
activate_last_pool();
}
else {
activate_next_pool();
}
return allocate(descriptor_set_layout);
}
return VKDescriptorSet(pool, vk_descriptor_set);
}
void VKDescriptorPools::free(VKDescriptorSet &descriptor_set)
{
VkDescriptorSet vk_descriptor_set = descriptor_set.vk_handle();
VkDescriptorPool vk_descriptor_pool = descriptor_set.vk_pool_handle();
BLI_assert(pools_.contains(vk_descriptor_pool));
vkFreeDescriptorSets(vk_device_, vk_descriptor_pool, 1, &vk_descriptor_set);
descriptor_set.mark_freed();
}
} // namespace blender::gpu

64
source/blender/gpu/vulkan/vk_descriptor_pools.hh Normal file
View File

@ -0,0 +1,64 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "BLI_vector.hh"
#include "vk_descriptor_set.hh"
namespace blender::gpu {
/**
* List of VkDescriptorPools.
*
* In Vulkan a pool is constructed with a fixed size per resource type. When more resources are
* needed it a next pool should be created. VKDescriptorPools will keep track of those pools and
* construct new pools when the previous one is exhausted.
*
* At the beginning of a new frame the descriptor pools are reset. This will start allocating
* again from the first descriptor pool in order to use freed space from previous pools.
*/
class VKDescriptorPools {
/**
* Pool sizes to use. When one descriptor pool is requested to allocate a descriptor but isn't
* able to do so, it will fail.
*
* Better defaults should be set later on, when we know more about our resource usage.
*/
static constexpr uint32_t POOL_SIZE_STORAGE_BUFFER = 1000;
static constexpr uint32_t POOL_SIZE_DESCRIPTOR_SETS = 1000;
static constexpr uint32_t POOL_SIZE_STORAGE_IMAGE = 1000;
static constexpr uint32_t POOL_SIZE_COMBINED_IMAGE_SAMPLER = 1000;
static constexpr uint32_t POOL_SIZE_UNIFORM_BUFFER = 1000;
VkDevice vk_device_ = VK_NULL_HANDLE;
Vector<VkDescriptorPool> pools_;
int64_t active_pool_index_ = 0;
public:
VKDescriptorPools();
~VKDescriptorPools();
void init(const VkDevice vk_device);
VKDescriptorSet allocate(const VkDescriptorSetLayout &descriptor_set_layout);
void free(VKDescriptorSet &descriptor_set);
/**
* Reset the pools to start looking for free space from the first descriptor pool.
*/
void reset();
private:
VkDescriptorPool active_pool_get();
void activate_next_pool();
void activate_last_pool();
bool is_last_pool_active();
void add_new_pool();
};
} // namespace blender::gpu

128
source/blender/gpu/vulkan/vk_descriptor_set.cc Normal file
View File

@ -0,0 +1,128 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_descriptor_set.hh"
#include "vk_index_buffer.hh"
#include "vk_storage_buffer.hh"
#include "vk_texture.hh"
#include "vk_vertex_buffer.hh"
#include "BLI_assert.h"
namespace blender::gpu {
VKDescriptorSet::~VKDescriptorSet()
{
if (vk_descriptor_set_ != VK_NULL_HANDLE) {
/* Handle should be given back to the pool.*/
VKContext &context = *VKContext::get();
context.descriptor_pools_get().free(*this);
BLI_assert(vk_descriptor_set_ == VK_NULL_HANDLE);
}
}
void VKDescriptorSet::mark_freed()
{
vk_descriptor_set_ = VK_NULL_HANDLE;
vk_descriptor_pool_ = VK_NULL_HANDLE;
}
void VKDescriptorSet::bind(VKStorageBuffer &buffer, const Location location)
{
Binding &binding = ensure_location(location);
binding.type = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
binding.vk_buffer = buffer.vk_handle();
binding.buffer_size = buffer.size_in_bytes();
}
void VKDescriptorSet::bind_as_ssbo(VKVertexBuffer &buffer, const Location location)
{
Binding &binding = ensure_location(location);
binding.type = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
binding.vk_buffer = buffer.vk_handle();
binding.buffer_size = buffer.size_used_get();
}
void VKDescriptorSet::bind_as_ssbo(VKIndexBuffer &buffer, const Location location)
{
Binding &binding = ensure_location(location);
binding.type = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
binding.vk_buffer = buffer.vk_handle();
binding.buffer_size = buffer.size_get();
}
void VKDescriptorSet::image_bind(VKTexture &texture, const Location location)
{
Binding &binding = ensure_location(location);
binding.type = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE;
binding.vk_image_view = texture.vk_image_view_handle();
}
VKDescriptorSet::Binding &VKDescriptorSet::ensure_location(const Location location)
{
for (Binding &binding : bindings_) {
if (binding.location == location) {
return binding;
}
}
Binding binding = {};
binding.location = location;
bindings_.append(binding);
return bindings_.last();
}
void VKDescriptorSet::update(VkDevice vk_device)
{
Vector<VkDescriptorBufferInfo> buffer_infos;
Vector<VkWriteDescriptorSet> descriptor_writes;
for (const Binding &binding : bindings_) {
if (!binding.is_buffer()) {
continue;
}
VkDescriptorBufferInfo buffer_info = {};
buffer_info.buffer = binding.vk_buffer;
buffer_info.range = binding.buffer_size;
buffer_infos.append(buffer_info);
VkWriteDescriptorSet write_descriptor = {};
write_descriptor.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET;
write_descriptor.dstSet = vk_descriptor_set_;
write_descriptor.dstBinding = binding.location;
write_descriptor.descriptorCount = 1;
write_descriptor.descriptorType = binding.type;
write_descriptor.pBufferInfo = &buffer_infos.last();
descriptor_writes.append(write_descriptor);
}
Vector<VkDescriptorImageInfo> image_infos;
for (const Binding &binding : bindings_) {
if (!binding.is_image()) {
continue;
}
VkDescriptorImageInfo image_info = {};
image_info.imageView = binding.vk_image_view;
image_info.imageLayout = VK_IMAGE_LAYOUT_GENERAL;
image_infos.append(image_info);
VkWriteDescriptorSet write_descriptor = {};
write_descriptor.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET;
write_descriptor.dstSet = vk_descriptor_set_;
write_descriptor.dstBinding = binding.location;
write_descriptor.descriptorCount = 1;
write_descriptor.descriptorType = binding.type;
write_descriptor.pImageInfo = &image_infos.last();
descriptor_writes.append(write_descriptor);
}
vkUpdateDescriptorSets(
vk_device, descriptor_writes.size(), descriptor_writes.data(), 0, nullptr);
bindings_.clear();
}
} // namespace blender::gpu

147
source/blender/gpu/vulkan/vk_descriptor_set.hh Normal file
View File

@ -0,0 +1,147 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "BLI_utility_mixins.hh"
#include "BLI_vector.hh"
#include "gpu_shader_private.hh"
#include "vk_common.hh"
namespace blender::gpu {
class VKStorageBuffer;
class VKVertexBuffer;
class VKIndexBuffer;
class VKTexture;
/**
* 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. Only for each #blender::gpu::shader::Frequency.
*/
class VKDescriptorSet : NonCopyable {
struct Binding;
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() = default;
public:
Location(const ShaderInput *shader_input) : binding(shader_input->location)
{
}
bool operator==(const Location &other) const
{
return binding == other.binding;
}
operator uint32_t() const
{
return binding;
}
friend struct Binding;
};
private:
struct Binding {
Location location;
VkDescriptorType type;
VkBuffer vk_buffer = VK_NULL_HANDLE;
VkDeviceSize buffer_size = 0;
VkImageView vk_image_view = VK_NULL_HANDLE;
Binding()
{
location.binding = 0;
}
bool is_buffer() const
{
return ELEM(type, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER);
}
bool is_image() const
{
return ELEM(type, VK_DESCRIPTOR_TYPE_STORAGE_IMAGE);
}
};
VkDescriptorPool vk_descriptor_pool_ = VK_NULL_HANDLE;
VkDescriptorSet vk_descriptor_set_ = VK_NULL_HANDLE;
/** A list of bindings that needs to be updated.*/
Vector<Binding> bindings_;
public:
VKDescriptorSet() = default;
VKDescriptorSet(VkDescriptorPool vk_descriptor_pool, VkDescriptorSet vk_descriptor_set)
: vk_descriptor_pool_(vk_descriptor_pool), vk_descriptor_set_(vk_descriptor_set)
{
}
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 bind_as_ssbo(VKVertexBuffer &buffer, Location location);
void bind_as_ssbo(VKIndexBuffer &buffer, Location location);
void bind(VKStorageBuffer &buffer, Location location);
void image_bind(VKTexture &texture, Location location);
/**
* Update the descriptor set on the device.
*/
void update(VkDevice vk_device);
void mark_freed();
private:
Binding &ensure_location(Location location);
};
} // namespace blender::gpu

35
source/blender/gpu/vulkan/vk_index_buffer.cc
View File

@ -6,6 +6,8 @@
*/
#include "vk_index_buffer.hh"
#include "vk_shader.hh"
#include "vk_shader_interface.hh"
namespace blender::gpu {
@ -13,12 +15,31 @@ void VKIndexBuffer::upload_data()
{
}
void VKIndexBuffer::bind_as_ssbo(uint /*binding*/)
void VKIndexBuffer::bind_as_ssbo(uint binding)
{
VKContext &context = *VKContext::get();
if (!buffer_.is_allocated()) {
allocate(context);
}
VKShader *shader = static_cast<VKShader *>(context.shader);
const VKShaderInterface &shader_interface = shader->interface_get();
const ShaderInput *shader_input = shader_interface.shader_input_get(
shader::ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER, binding);
shader->pipeline_get().descriptor_set_get().bind_as_ssbo(*this, shader_input);
}
void VKIndexBuffer::read(uint32_t * /*data*/) const
void VKIndexBuffer::read(uint32_t *data) const
{
VKContext &context = *VKContext::get();
VKCommandBuffer &command_buffer = context.command_buffer_get();
command_buffer.submit();
void *mapped_memory;
if (buffer_.map(context, &mapped_memory)) {
memcpy(data, mapped_memory, size_get());
buffer_.unmap(context);
}
}
void VKIndexBuffer::update_sub(uint /*start*/, uint /*len*/, const void * /*data*/)
@ -29,4 +50,14 @@ void VKIndexBuffer::strip_restart_indices()
{
}
void VKIndexBuffer::allocate(VKContext &context)
{
GPUUsageType usage = data_ == nullptr ? GPU_USAGE_DEVICE_ONLY : GPU_USAGE_STATIC;
buffer_.create(context,
size_get(),
usage,
static_cast<VkBufferUsageFlagBits>(VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
VK_BUFFER_USAGE_INDEX_BUFFER_BIT));
}
} // namespace blender::gpu

10
source/blender/gpu/vulkan/vk_index_buffer.hh
View File

@ -9,9 +9,13 @@
#include "gpu_index_buffer_private.hh"
#include "vk_buffer.hh"
namespace blender::gpu {
class VKIndexBuffer : public IndexBuf {
VKBuffer buffer_;
public:
void upload_data() override;
@ -21,8 +25,14 @@ class VKIndexBuffer : public IndexBuf {
void update_sub(uint start, uint len, const void *data) override;
VkBuffer vk_handle()
{
return buffer_.vk_handle();
}
private:
void strip_restart_indices() override;
void allocate(VKContext &context);
};
} // namespace blender::gpu

4
source/blender/gpu/vulkan/vk_memory.cc
View File

@ -37,7 +37,9 @@ void *vk_memory_reallocation(void *user_data,
void vk_memory_free(void * /*user_data*/, void *memory)
{
MEM_freeN(memory);
if (memory != nullptr) {
MEM_freeN(memory);
}
}
#endif

6
source/blender/gpu/vulkan/vk_memory.hh
View File

@ -7,11 +7,7 @@
#pragma once
#ifdef __APPLE__
# include <MoltenVK/vk_mvk_moltenvk.h>
#else
# include <vulkan/vulkan.h>
#endif
#include "vk_common.hh"
namespace blender::gpu {

68
source/blender/gpu/vulkan/vk_pipeline.cc Normal file
View File

@ -0,0 +1,68 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_pipeline.hh"
#include "vk_context.hh"
#include "vk_memory.hh"
namespace blender::gpu {
VKPipeline::VKPipeline(VkPipeline vk_pipeline, VKDescriptorSet &&vk_descriptor_set)
: vk_pipeline_(vk_pipeline)
{
descriptor_set_ = std::move(vk_descriptor_set);
}
VKPipeline::~VKPipeline()
{
VK_ALLOCATION_CALLBACKS
VkDevice vk_device = VKContext::get()->device_get();
if (vk_pipeline_ != VK_NULL_HANDLE) {
vkDestroyPipeline(vk_device, vk_pipeline_, vk_allocation_callbacks);
}
}
VKPipeline VKPipeline::create_compute_pipeline(VKContext &context,
VkShaderModule compute_module,
VkDescriptorSetLayout &descriptor_set_layout,
VkPipelineLayout &pipeline_layout)
{
VK_ALLOCATION_CALLBACKS
VkDevice vk_device = context.device_get();
VkComputePipelineCreateInfo pipeline_info = {};
pipeline_info.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO;
pipeline_info.flags = 0;
pipeline_info.stage = {};
pipeline_info.stage.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
pipeline_info.stage.flags = 0;
pipeline_info.stage.stage = VK_SHADER_STAGE_COMPUTE_BIT;
pipeline_info.stage.module = compute_module;
pipeline_info.layout = pipeline_layout;
pipeline_info.stage.pName = "main";
VkPipeline pipeline;
if (vkCreateComputePipelines(
vk_device, nullptr, 1, &pipeline_info, vk_allocation_callbacks, &pipeline) !=
VK_SUCCESS) {
return VKPipeline();
}
VKDescriptorSet descriptor_set = context.descriptor_pools_get().allocate(descriptor_set_layout);
return VKPipeline(pipeline, std::move(descriptor_set));
}
VkPipeline VKPipeline::vk_handle() const
{
return vk_pipeline_;
}
bool VKPipeline::is_valid() const
{
return vk_pipeline_ != VK_NULL_HANDLE;
}
} // namespace blender::gpu

50
source/blender/gpu/vulkan/vk_pipeline.hh Normal file
View File

@ -0,0 +1,50 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "BLI_utility_mixins.hh"
#include "BLI_vector.hh"
#include "vk_common.hh"
#include "vk_descriptor_set.hh"
namespace blender::gpu {
class VKContext;
class VKPipeline : NonCopyable {
VKDescriptorSet descriptor_set_;
VkPipeline vk_pipeline_ = VK_NULL_HANDLE;
public:
VKPipeline() = default;
virtual ~VKPipeline();
VKPipeline(VkPipeline vk_pipeline, VKDescriptorSet &&vk_descriptor_set);
VKPipeline &operator=(VKPipeline &&other)
{
vk_pipeline_ = other.vk_pipeline_;
other.vk_pipeline_ = VK_NULL_HANDLE;
descriptor_set_ = std::move(other.descriptor_set_);
return *this;
}
static VKPipeline create_compute_pipeline(VKContext &context,
VkShaderModule compute_module,
VkDescriptorSetLayout &descriptor_set_layout,
VkPipelineLayout &pipeline_layouts);
VKDescriptorSet &descriptor_set_get()
{
return descriptor_set_;
}
VkPipeline vk_handle() const;
bool is_valid() const;
};
} // namespace blender::gpu

255
source/blender/gpu/vulkan/vk_shader.cc
View File

@ -9,6 +9,7 @@
#include "vk_backend.hh"
#include "vk_memory.hh"
#include "vk_shader_interface.hh"
#include "vk_shader_log.hh"
#include "BLI_string_utils.h"
@ -324,9 +325,11 @@ static std::ostream &print_qualifier(std::ostream &os, const Qualifier &qualifie
return os;
}
static void print_resource(std::ostream &os, const ShaderCreateInfo::Resource &res)
static void print_resource(std::ostream &os,
const ShaderInput &shader_input,
const ShaderCreateInfo::Resource &res)
{
os << "layout(binding = " << res.slot;
os << "layout(binding = " << shader_input.location;
if (res.bind_type == ShaderCreateInfo::Resource::BindType::IMAGE) {
os << ", " << to_string(res.image.format);
}
@ -372,6 +375,18 @@ static void print_resource(std::ostream &os, const ShaderCreateInfo::Resource &r
}
}
static void print_resource(std::ostream &os,
const VKShaderInterface &shader_interface,
const ShaderCreateInfo::Resource &res)
{
const ShaderInput *shader_input = shader_interface.shader_input_get(res);
if (shader_input == nullptr) {
BLI_assert_msg(shader_input, "Cannot find shader input for resource");
return;
}
print_resource(os, *shader_input, res);
}
static void print_resource_alias(std::ostream &os, const ShaderCreateInfo::Resource &res)
{
int64_t array_offset;
@ -603,6 +618,14 @@ VKShader::~VKShader()
vkDestroyShaderModule(device, compute_module_, vk_allocation_callbacks);
compute_module_ = VK_NULL_HANDLE;
}
if (pipeline_layout_ != VK_NULL_HANDLE) {
vkDestroyPipelineLayout(device, pipeline_layout_, vk_allocation_callbacks);
pipeline_layout_ = VK_NULL_HANDLE;
}
if (layout_ != VK_NULL_HANDLE) {
vkDestroyDescriptorSetLayout(device, layout_, vk_allocation_callbacks);
layout_ = VK_NULL_HANDLE;
}
}
void VKShader::build_shader_module(MutableSpan<const char *> sources,
@ -650,52 +673,181 @@ bool VKShader::finalize(const shader::ShaderCreateInfo *info)
return false;
}
if (vertex_module_ != VK_NULL_HANDLE) {
VKShaderInterface *vk_interface = new VKShaderInterface();
vk_interface->init(*info);
VkDevice vk_device = context_->device_get();
if (!finalize_descriptor_set_layouts(vk_device, *vk_interface, *info)) {
return false;
}
if (!finalize_pipeline_layout(vk_device, *info)) {
return false;
}
/* TODO we might need to move the actual pipeline construction to a later stage as the graphics
* pipeline requires more data before it can be constructed.*/
bool result;
if (is_graphics_shader()) {
BLI_assert((fragment_module_ != VK_NULL_HANDLE && info->tf_type_ == GPU_SHADER_TFB_NONE) ||
(fragment_module_ == VK_NULL_HANDLE && info->tf_type_ != GPU_SHADER_TFB_NONE));
BLI_assert(compute_module_ == VK_NULL_HANDLE);
VkPipelineShaderStageCreateInfo vertex_stage_info = {};
vertex_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
vertex_stage_info.stage = VK_SHADER_STAGE_VERTEX_BIT;
vertex_stage_info.module = vertex_module_;
vertex_stage_info.pName = "main";
pipeline_infos_.append(vertex_stage_info);
if (geometry_module_ != VK_NULL_HANDLE) {
VkPipelineShaderStageCreateInfo geo_stage_info = {};
geo_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
geo_stage_info.stage = VK_SHADER_STAGE_GEOMETRY_BIT;
geo_stage_info.module = geometry_module_;
geo_stage_info.pName = "main";
pipeline_infos_.append(geo_stage_info);
}
if (fragment_module_ != VK_NULL_HANDLE) {
VkPipelineShaderStageCreateInfo fragment_stage_info = {};
fragment_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
fragment_stage_info.stage = VK_SHADER_STAGE_FRAGMENT_BIT;
fragment_stage_info.module = fragment_module_;
fragment_stage_info.pName = "main";
pipeline_infos_.append(fragment_stage_info);
}
result = finalize_graphics_pipeline(vk_device);
}
else {
BLI_assert(vertex_module_ == VK_NULL_HANDLE);
BLI_assert(geometry_module_ == VK_NULL_HANDLE);
BLI_assert(fragment_module_ == VK_NULL_HANDLE);
BLI_assert(compute_module_ != VK_NULL_HANDLE);
VkPipelineShaderStageCreateInfo compute_stage_info = {};
compute_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
compute_stage_info.stage = VK_SHADER_STAGE_GEOMETRY_BIT;
compute_stage_info.module = compute_module_;
compute_stage_info.pName = "main";
pipeline_infos_.append(compute_stage_info);
compute_pipeline_ = VKPipeline::create_compute_pipeline(
*context_, compute_module_, layout_, pipeline_layout_);
result = compute_pipeline_.is_valid();
}
#ifdef NDEBUG
UNUSED_VARS(info);
#endif
if (result) {
interface = vk_interface;
}
else {
delete vk_interface;
}
return result;
}
bool VKShader::finalize_graphics_pipeline(VkDevice /*vk_device */)
{
Vector<VkPipelineShaderStageCreateInfo> pipeline_stages;
VkPipelineShaderStageCreateInfo vertex_stage_info = {};
vertex_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
vertex_stage_info.stage = VK_SHADER_STAGE_VERTEX_BIT;
vertex_stage_info.module = vertex_module_;
vertex_stage_info.pName = "main";
pipeline_stages.append(vertex_stage_info);
if (geometry_module_ != VK_NULL_HANDLE) {
VkPipelineShaderStageCreateInfo geo_stage_info = {};
geo_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
geo_stage_info.stage = VK_SHADER_STAGE_GEOMETRY_BIT;
geo_stage_info.module = geometry_module_;
geo_stage_info.pName = "main";
pipeline_stages.append(geo_stage_info);
}
if (fragment_module_ != VK_NULL_HANDLE) {
VkPipelineShaderStageCreateInfo fragment_stage_info = {};
fragment_stage_info.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
fragment_stage_info.stage = VK_SHADER_STAGE_FRAGMENT_BIT;
fragment_stage_info.module = fragment_module_;
fragment_stage_info.pName = "main";
pipeline_stages.append(fragment_stage_info);
}
return true;
}
bool VKShader::finalize_pipeline_layout(VkDevice vk_device,
const shader::ShaderCreateInfo & /*info*/)
{
VK_ALLOCATION_CALLBACKS
const uint32_t layout_count = layout_ == VK_NULL_HANDLE ? 0 : 1;
VkPipelineLayoutCreateInfo pipeline_info = {};
pipeline_info.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO;
pipeline_info.flags = 0;
pipeline_info.setLayoutCount = layout_count;
pipeline_info.pSetLayouts = &layout_;
if (vkCreatePipelineLayout(
vk_device, &pipeline_info, vk_allocation_callbacks, &pipeline_layout_) != VK_SUCCESS) {
return false;
};
return true;
}
static VkDescriptorType descriptor_type(
const shader::ShaderCreateInfo::Resource::BindType bind_type)
{
switch (bind_type) {
case shader::ShaderCreateInfo::Resource::BindType::IMAGE:
return VK_DESCRIPTOR_TYPE_STORAGE_IMAGE;
case shader::ShaderCreateInfo::Resource::BindType::SAMPLER:
return VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER;
case shader::ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER:
return VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
case shader::ShaderCreateInfo::Resource::BindType::UNIFORM_BUFFER:
return VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER;
}
BLI_assert_unreachable();
return VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER;
}
static VkDescriptorSetLayoutBinding create_descriptor_set_layout_binding(
const ShaderInput &shader_input, const shader::ShaderCreateInfo::Resource &resource)
{
VkDescriptorSetLayoutBinding binding = {};
binding.binding = shader_input.location;
binding.descriptorType = descriptor_type(resource.bind_type);
binding.descriptorCount = 1;
binding.stageFlags = VK_SHADER_STAGE_ALL;
binding.pImmutableSamplers = nullptr;
return binding;
}
static void add_descriptor_set_layout_bindings(
const VKShaderInterface &interface,
const Vector<shader::ShaderCreateInfo::Resource> &resources,
Vector<VkDescriptorSetLayoutBinding> &r_bindings)
{
for (const shader::ShaderCreateInfo::Resource &resource : resources) {
const ShaderInput *shader_input = interface.shader_input_get(resource);
if (shader_input == nullptr) {
BLI_assert_msg(shader_input, "Cannot find shader input for resource.");
continue;
}
r_bindings.append(create_descriptor_set_layout_binding(*shader_input, resource));
}
}
static VkDescriptorSetLayoutCreateInfo create_descriptor_set_layout(
const VKShaderInterface &interface,
const Vector<shader::ShaderCreateInfo::Resource> &resources,
Vector<VkDescriptorSetLayoutBinding> &r_bindings)
{
add_descriptor_set_layout_bindings(interface, resources, r_bindings);
VkDescriptorSetLayoutCreateInfo set_info = {};
set_info.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO;
set_info.flags = 0;
set_info.pNext = nullptr;
set_info.bindingCount = r_bindings.size();
set_info.pBindings = r_bindings.data();
return set_info;
}
bool VKShader::finalize_descriptor_set_layouts(VkDevice vk_device,
const VKShaderInterface &shader_interface,
const shader::ShaderCreateInfo &info)
{
if (info.pass_resources_.is_empty() && info.batch_resources_.is_empty()) {
return true;
}
VK_ALLOCATION_CALLBACKS
/* Currently we create a single descriptor set. The goal would be to create one descriptor set
* for Frequency::PASS/BATCH. This isn't possible as areas expect that the binding location is
* static and predictable (eevee-next) or the binding location can be mapped to a single number
* (python). */
Vector<ShaderCreateInfo::Resource> all_resources;
all_resources.extend(info.pass_resources_);
all_resources.extend(info.batch_resources_);
Vector<VkDescriptorSetLayoutBinding> bindings;
VkDescriptorSetLayoutCreateInfo layout_info = create_descriptor_set_layout(
shader_interface, all_resources, bindings);
if (vkCreateDescriptorSetLayout(vk_device, &layout_info, vk_allocation_callbacks, &layout_) !=
VK_SUCCESS) {
return false;
};
return true;
}
@ -716,10 +868,23 @@ void VKShader::transform_feedback_disable()
void VKShader::bind()
{
VKContext *context = VKContext::get();
if (is_compute_shader()) {
context->command_buffer_get().bind(compute_pipeline_, VK_PIPELINE_BIND_POINT_COMPUTE);
}
else {
BLI_assert_unreachable();
}
}
void VKShader::unbind()
{
if (is_compute_shader()) {
}
else {
BLI_assert_unreachable();
}
}
void VKShader::uniform_float(int /*location*/,
@ -737,11 +902,13 @@ void VKShader::uniform_int(int /*location*/,
std::string VKShader::resources_declare(const shader::ShaderCreateInfo &info) const
{
VKShaderInterface interface;
interface.init(info);
std::stringstream ss;
ss << "\n/* Pass Resources. */\n";
for (const ShaderCreateInfo::Resource &res : info.pass_resources_) {
print_resource(ss, res);
print_resource(ss, interface, res);
}
for (const ShaderCreateInfo::Resource &res : info.pass_resources_) {
print_resource_alias(ss, res);
@ -749,7 +916,7 @@ std::string VKShader::resources_declare(const shader::ShaderCreateInfo &info) co
ss << "\n/* Batch Resources. */\n";
for (const ShaderCreateInfo::Resource &res : info.batch_resources_) {
print_resource(ss, res);
print_resource(ss, interface, res);
}
for (const ShaderCreateInfo::Resource &res : info.batch_resources_) {
print_resource_alias(ss, res);
@ -958,4 +1125,14 @@ int VKShader::program_handle_get() const
return -1;
}
VKPipeline &VKShader::pipeline_get()
{
return compute_pipeline_;
}
const VKShaderInterface &VKShader::interface_get() const
{
return *static_cast<const VKShaderInterface *>(interface);
}
} // namespace blender::gpu

28
source/blender/gpu/vulkan/vk_shader.hh
View File

@ -15,6 +15,7 @@
#include "BLI_string_ref.hh"
namespace blender::gpu {
class VKShaderInterface;
class VKShader : public Shader {
private:
@ -24,7 +25,9 @@ class VKShader : public Shader {
VkShaderModule fragment_module_ = VK_NULL_HANDLE;
VkShaderModule compute_module_ = VK_NULL_HANDLE;
bool compilation_failed_ = false;
Vector<VkPipelineShaderStageCreateInfo> pipeline_infos_;
VkDescriptorSetLayout layout_ = VK_NULL_HANDLE;
VkPipelineLayout pipeline_layout_ = VK_NULL_HANDLE;
VKPipeline compute_pipeline_;
public:
VKShader(const char *name);
@ -58,12 +61,35 @@ class VKShader : public Shader {
/* DEPRECATED: Kept only because of BGL API. */
int program_handle_get() const override;
VKPipeline &pipeline_get();
VkPipelineLayout vk_pipeline_layout_get() const
{
return pipeline_layout_;
}
const VKShaderInterface &interface_get() const;
private:
Vector<uint32_t> compile_glsl_to_spirv(Span<const char *> sources, shaderc_shader_kind kind);
void build_shader_module(Span<uint32_t> spirv_module, VkShaderModule *r_shader_module);
void build_shader_module(MutableSpan<const char *> sources,
shaderc_shader_kind stage,
VkShaderModule *r_shader_module);
bool finalize_descriptor_set_layouts(VkDevice vk_device,
const VKShaderInterface &shader_interface,
const shader::ShaderCreateInfo &info);
bool finalize_pipeline_layout(VkDevice vk_device, const shader::ShaderCreateInfo &info);
bool finalize_graphics_pipeline(VkDevice vk_device);
bool is_graphics_shader() const
{
return !is_compute_shader();
}
bool is_compute_shader() const
{
return compute_module_ != VK_NULL_HANDLE;
}
};
} // namespace blender::gpu

117
source/blender/gpu/vulkan/vk_shader_interface.cc Normal file
View File

@ -0,0 +1,117 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_shader_interface.hh"
namespace blender::gpu {
void VKShaderInterface::init(const shader::ShaderCreateInfo &info)
{
using namespace blender::gpu::shader;
attr_len_ = 0;
uniform_len_ = 0;
ssbo_len_ = 0;
ubo_len_ = 0;
image_offset_ = -1;
Vector<ShaderCreateInfo::Resource> all_resources;
all_resources.extend(info.pass_resources_);
all_resources.extend(info.batch_resources_);
for (ShaderCreateInfo::Resource &res : all_resources) {
switch (res.bind_type) {
case ShaderCreateInfo::Resource::BindType::IMAGE:
uniform_len_++;
break;
case ShaderCreateInfo::Resource::BindType::SAMPLER:
image_offset_ = max_ii(image_offset_, res.slot);
uniform_len_++;
break;
case ShaderCreateInfo::Resource::BindType::UNIFORM_BUFFER:
ubo_len_++;
break;
case ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER:
ssbo_len_++;
break;
}
}
/* Make sure that the image slots don't overlap with the sampler slots.*/
image_offset_ += 1;
int32_t input_tot_len = ubo_len_ + uniform_len_ + ssbo_len_;
inputs_ = static_cast<ShaderInput *>(
MEM_calloc_arrayN(input_tot_len, sizeof(ShaderInput), __func__));
ShaderInput *input = inputs_;
name_buffer_ = (char *)MEM_mallocN(info.interface_names_size_, "name_buffer");
uint32_t name_buffer_offset = 0;
int location = 0;
/* Uniform blocks */
for (const ShaderCreateInfo::Resource &res : all_resources) {
if (res.bind_type == ShaderCreateInfo::Resource::BindType::UNIFORM_BUFFER) {
copy_input_name(input, res.image.name, name_buffer_, name_buffer_offset);
input->location = location++;
input->binding = res.slot;
input++;
}
}
/* Images, Samplers and buffers. */
for (const ShaderCreateInfo::Resource &res : all_resources) {
if (res.bind_type == ShaderCreateInfo::Resource::BindType::SAMPLER) {
copy_input_name(input, res.sampler.name, name_buffer_, name_buffer_offset);
input->location = location++;
input->binding = res.slot;
input++;
}
else if (res.bind_type == ShaderCreateInfo::Resource::BindType::IMAGE) {
copy_input_name(input, res.image.name, name_buffer_, name_buffer_offset);
input->location = location++;
input->binding = res.slot + image_offset_;
input++;
}
}
/* Storage buffers */
for (const ShaderCreateInfo::Resource &res : all_resources) {
if (res.bind_type == ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER) {
copy_input_name(input, res.storagebuf.name, name_buffer_, name_buffer_offset);
input->location = location++;
input->binding = res.slot;
input++;
}
}
sort_inputs();
}
const ShaderInput *VKShaderInterface::shader_input_get(
const shader::ShaderCreateInfo::Resource &resource) const
{
return shader_input_get(resource.bind_type, resource.slot);
}
const ShaderInput *VKShaderInterface::shader_input_get(
const shader::ShaderCreateInfo::Resource::BindType &bind_type, int binding) const
{
switch (bind_type) {
case shader::ShaderCreateInfo::Resource::BindType::IMAGE:
return texture_get(binding + image_offset_);
case shader::ShaderCreateInfo::Resource::BindType::SAMPLER:
return texture_get(binding);
case shader::ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER:
return ssbo_get(binding);
case shader::ShaderCreateInfo::Resource::BindType::UNIFORM_BUFFER:
return ubo_get(binding);
}
return nullptr;
}
} // namespace blender::gpu

39
source/blender/gpu/vulkan/vk_shader_interface.hh Normal file
View File

@ -0,0 +1,39 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "gpu_shader_create_info.hh"
#include "gpu_shader_interface.hh"
namespace blender::gpu {
class VKShaderInterface : public ShaderInterface {
private:
/**
* Offset when searching for a shader input based on a binding number.
*
* When shaders combine images and samplers, the images have to be offset to find the correct
* shader input. Both textures and images are stored in the uniform list and their ID can be
* overlapping.
*/
uint32_t image_offset_ = 0;
public:
VKShaderInterface() = default;
void init(const shader::ShaderCreateInfo &info);
/**
* Retrieve the shader input for the given resource.
*
* nullptr is returned when resource could not be found.
* Should only happen when still developing the Vulkan shader.
*/
const ShaderInput *shader_input_get(const shader::ShaderCreateInfo::Resource &resource) const;
const ShaderInput *shader_input_get(
const shader::ShaderCreateInfo::Resource::BindType &bind_type, int binding) const;
};
} // namespace blender::gpu

59
source/blender/gpu/vulkan/vk_state_manager.cc Normal file
View File

@ -0,0 +1,59 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#include "vk_state_manager.hh"
#include "vk_texture.hh"
namespace blender::gpu {
void VKStateManager::apply_state()
{
}
void VKStateManager::force_state()
{
}
void VKStateManager::issue_barrier(eGPUBarrier /*barrier_bits*/)
{
VKContext &context = *VKContext::get();
VKCommandBuffer &command_buffer = context.command_buffer_get();
/* TODO: Pipeline barriers should be added. We might be able to extract it from
* the actual pipeline, later on, but for now we submit the work as barrier. */
command_buffer.submit();
}
void VKStateManager::texture_bind(Texture * /*tex*/, eGPUSamplerState /*sampler*/, int /*unit*/)
{
}
void VKStateManager::texture_unbind(Texture * /*tex*/)
{
}
void VKStateManager::texture_unbind_all()
{
}
void VKStateManager::image_bind(Texture *tex, int binding)
{
VKTexture *texture = unwrap(tex);
texture->image_bind(binding);
}
void VKStateManager::image_unbind(Texture * /*tex*/)
{
}
void VKStateManager::image_unbind_all()
{
}
void VKStateManager::texture_unpack_row_length_set(uint /*len*/)
{
}
} // namespace blender::gpu

30
source/blender/gpu/vulkan/vk_state_manager.hh Normal file
View File

@ -0,0 +1,30 @@
/* SPDX-License-Identifier: GPL-2.0-or-later
* Copyright 2023 Blender Foundation. All rights reserved. */
/** \file
* \ingroup gpu
*/
#pragma once
#include "gpu_state_private.hh"
namespace blender::gpu {
class VKStateManager : public StateManager {
public:
void apply_state() override;
void force_state() override;
void issue_barrier(eGPUBarrier barrier_bits) override;
void texture_bind(Texture *tex, eGPUSamplerState sampler, int unit) override;
void texture_unbind(Texture *tex) override;
void texture_unbind_all() override;
void image_bind(Texture *tex, int unit) override;
void image_unbind(Texture *tex) override;
void image_unbind_all() override;
void texture_unpack_row_length_set(uint len) override;
};
} // namespace blender::gpu

41
source/blender/gpu/vulkan/vk_storage_buffer.cc
View File

@ -4,19 +4,39 @@
/** \file
* \ingroup gpu
*/
#include "vk_shader.hh"
#include "vk_shader_interface.hh"
#include "vk_vertex_buffer.hh"
#include "vk_storage_buffer.hh"
namespace blender::gpu {
void VKStorageBuffer::update(const void * /*data*/)
void VKStorageBuffer::update(const void *data)
{
VKContext &context = *VKContext::get();
if (!buffer_.is_allocated()) {
allocate(context);
}
buffer_.update(context, data);
}
void VKStorageBuffer::bind(int /*slot*/)
void VKStorageBuffer::allocate(VKContext &context)
{
buffer_.create(context, size_in_bytes_, usage_, VK_BUFFER_USAGE_STORAGE_BUFFER_BIT);
}
void VKStorageBuffer::bind(int slot)
{
VKContext &context = *VKContext::get();
if (!buffer_.is_allocated()) {
allocate(context);
}
VKShader *shader = static_cast<VKShader *>(context.shader);
const VKShaderInterface &shader_interface = shader->interface_get();
const ShaderInput *shader_input = shader_interface.shader_input_get(
shader::ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER, slot);
shader->pipeline_get().descriptor_set_get().bind(*this, shader_input);
}
void VKStorageBuffer::unbind()
@ -35,8 +55,21 @@ void VKStorageBuffer::copy_sub(VertBuf * /*src*/,
{
}
void VKStorageBuffer::read(void * /*data*/)
void VKStorageBuffer::read(void *data)
{
VKContext &context = *VKContext::get();
if (!buffer_.is_allocated()) {
allocate(context);
}
VKCommandBuffer &command_buffer = context.command_buffer_get();
command_buffer.submit();
void *mapped_memory;
if (buffer_.map(context, &mapped_memory)) {
memcpy(data, mapped_memory, size_in_bytes_);
buffer_.unmap(context);
}
}
} // namespace blender::gpu

23
source/blender/gpu/vulkan/vk_storage_buffer.hh
View File

@ -10,12 +10,20 @@
#include "GPU_texture.h"
#include "gpu_storage_buffer_private.hh"
#include "gpu_vertex_buffer_private.hh"
#include "vk_buffer.hh"
namespace blender::gpu {
class VertBuf;
class VKStorageBuffer : public StorageBuf {
GPUUsageType usage_;
VKBuffer buffer_;
public:
VKStorageBuffer(int size, const char *name) : StorageBuf(size, name)
VKStorageBuffer(int size, GPUUsageType usage, const char *name)
: StorageBuf(size, name), usage_(usage)
{
}
@ -25,6 +33,19 @@ class VKStorageBuffer : public StorageBuf {
void clear(eGPUTextureFormat internal_format, eGPUDataFormat data_format, void *data) override;
void copy_sub(VertBuf *src, uint dst_offset, uint src_offset, uint copy_size) override;
void read(void *data) override;
VkBuffer vk_handle() const
{
return buffer_.vk_handle();
}
int64_t size_in_bytes() const
{
return buffer_.size_in_bytes();
}
private:
void allocate(VKContext &context);
};
} // namespace blender::gpu

164
source/blender/gpu/vulkan/vk_texture.cc
View File

@ -7,8 +7,25 @@
#include "vk_texture.hh"
#include "vk_buffer.hh"
#include "vk_context.hh"
#include "vk_memory.hh"
#include "vk_shader.hh"
#include "vk_shader_interface.hh"
#include "BKE_global.h"
namespace blender::gpu {
VKTexture::~VKTexture()
{
VK_ALLOCATION_CALLBACKS
VKContext &context = *VKContext::get();
vmaDestroyImage(context.mem_allocator_get(), vk_image_, allocation_);
vkDestroyImageView(context.device_get(), vk_image_view_, vk_allocation_callbacks);
}
void VKTexture::generate_mipmap()
{
}
@ -33,9 +50,49 @@ void VKTexture::mip_range_set(int /*min*/, int /*max*/)
{
}
void *VKTexture::read(int /*mip*/, eGPUDataFormat /*format*/)
void *VKTexture::read(int mip, eGPUDataFormat format)
{
return nullptr;
/* Vulkan images cannot be directly mapped to host memory and requires a staging buffer.*/
VKContext &context = *VKContext::get();
VKBuffer staging_buffer;
/* NOTE: mip_size_get() won't override any dimension that is equal to 0. */
int extent[3] = {1, 1, 1};
mip_size_get(mip, extent);
size_t sample_len = extent[0] * extent[1] * extent[2];
/* NOTE: to_bytesize returns number of bits. */
size_t device_memory_size = sample_len * to_component_len(format_) * to_bytesize(format_) / 8;
/* NOTE: to_bytesize returns number of bytes here. */
size_t host_memory_size = sample_len * to_bytesize(format_, format);
staging_buffer.create(
context, device_memory_size, GPU_USAGE_DEVICE_ONLY, VK_BUFFER_USAGE_TRANSFER_DST_BIT);
VkBufferImageCopy region = {};
region.imageExtent.width = extent[0];
region.imageExtent.height = extent[1];
region.imageExtent.depth = extent[2];
region.imageSubresource.aspectMask = to_vk_image_aspect_flag_bits(format_);
region.imageSubresource.mipLevel = mip;
region.imageSubresource.layerCount = 1;
VKCommandBuffer &command_buffer = context.command_buffer_get();
command_buffer.copy(staging_buffer, *this, Span<VkBufferImageCopy>(&region, 1));
command_buffer.submit();
void *mapped_data;
staging_buffer.map(context, &mapped_data);
void *data = MEM_mallocN(host_memory_size, __func__);
/* TODO: add conversion when data format is different.*/
BLI_assert_msg(device_memory_size == host_memory_size,
"Memory data conversions not implemented yet");
memcpy(data, mapped_data, host_memory_size);
staging_buffer.unmap(context);
return data;
}
void VKTexture::update_sub(int /*mip*/,
@ -61,7 +118,10 @@ uint VKTexture::gl_bindcode_get() const
bool VKTexture::init_internal()
{
return false;
/* Initialization can only happen after the usage is known. By the current API this isn't set
* at this moment, so we cannot initialize here. The initialization is postponed until the
* allocation of the texture on the device.*/
return true;
}
bool VKTexture::init_internal(GPUVertBuf * /*vbo*/)
@ -74,4 +134,102 @@ bool VKTexture::init_internal(const GPUTexture * /*src*/, int /*mip_offset*/, in
return false;
}
bool VKTexture::is_allocated()
{
return vk_image_ != VK_NULL_HANDLE && allocation_ != VK_NULL_HANDLE;
}
bool VKTexture::allocate()
{
BLI_assert(!is_allocated());
int extent[3] = {1, 1, 1};
mip_size_get(0, extent);
VKContext &context = *VKContext::get();
VkImageCreateInfo image_info = {};
image_info.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO;
image_info.imageType = to_vk_image_type(type_);
image_info.extent.width = extent[0];
image_info.extent.height = extent[1];
image_info.extent.depth = extent[2];
image_info.mipLevels = 1;
image_info.arrayLayers = 1;
image_info.format = to_vk_format(format_);
image_info.tiling = VK_IMAGE_TILING_LINEAR;
image_info.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED;
image_info.usage = VK_IMAGE_USAGE_TRANSFER_SRC_BIT | VK_IMAGE_USAGE_SAMPLED_BIT |
VK_IMAGE_USAGE_STORAGE_BIT;
image_info.samples = VK_SAMPLE_COUNT_1_BIT;
VkResult result;
if (G.debug &= G_DEBUG_GPU) {
VkImageFormatProperties image_format = {};
result = vkGetPhysicalDeviceImageFormatProperties(context.physical_device_get(),
image_info.format,
image_info.imageType,
image_info.tiling,
image_info.usage,
image_info.flags,
&image_format);
if (result != VK_SUCCESS) {
printf("Image type not supported on device.\n");
return false;
}
}
VmaAllocationCreateInfo allocCreateInfo = {};
allocCreateInfo.usage = VMA_MEMORY_USAGE_AUTO;
allocCreateInfo.flags = static_cast<VmaAllocationCreateFlagBits>(
VMA_ALLOCATION_CREATE_HOST_ACCESS_RANDOM_BIT | VMA_ALLOCATION_CREATE_MAPPED_BIT);
allocCreateInfo.priority = 1.0f;
result = vmaCreateImage(context.mem_allocator_get(),
&image_info,
&allocCreateInfo,
&vk_image_,
&allocation_,
nullptr);
if (result != VK_SUCCESS) {
return false;
}
/* Promote image to the correct layout.*/
VkImageMemoryBarrier barrier{};
barrier.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER;
barrier.oldLayout = VK_IMAGE_LAYOUT_UNDEFINED;
barrier.newLayout = VK_IMAGE_LAYOUT_GENERAL;
barrier.image = vk_image_;
barrier.subresourceRange.aspectMask = to_vk_image_aspect_flag_bits(format_);
barrier.subresourceRange.levelCount = VK_REMAINING_MIP_LEVELS;
barrier.subresourceRange.layerCount = VK_REMAINING_ARRAY_LAYERS;
context.command_buffer_get().pipeline_barrier(Span<VkImageMemoryBarrier>(&barrier, 1));
VK_ALLOCATION_CALLBACKS
VkImageViewCreateInfo image_view_info = {};
image_view_info.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO;
image_view_info.image = vk_image_;
image_view_info.viewType = to_vk_image_view_type(type_);
image_view_info.format = to_vk_format(format_);
image_view_info.components = to_vk_component_mapping(format_);
image_view_info.subresourceRange.aspectMask = to_vk_image_aspect_flag_bits(format_);
image_view_info.subresourceRange.levelCount = VK_REMAINING_MIP_LEVELS;
image_view_info.subresourceRange.layerCount = VK_REMAINING_ARRAY_LAYERS;
result = vkCreateImageView(
context.device_get(), &image_view_info, vk_allocation_callbacks, &vk_image_view_);
return result == VK_SUCCESS;
}
void VKTexture::image_bind(int binding)
{
if (!is_allocated()) {
allocate();
}
VKContext &context = *VKContext::get();
VKShader *shader = static_cast<VKShader *>(context.shader);
VKDescriptorSet::Location location(shader->interface_get().shader_input_get(
shader::ShaderCreateInfo::Resource::BindType::IMAGE, binding));
shader->pipeline_get().descriptor_set_get().image_bind(*this, location);
}
} // namespace blender::gpu

34
source/blender/gpu/vulkan/vk_texture.hh
View File

@ -8,14 +8,22 @@
#pragma once
#include "gpu_texture_private.hh"
#include "vk_context.hh"
#include "vk_mem_alloc.h"
namespace blender::gpu {
class VKTexture : public Texture {
VkImage vk_image_ = VK_NULL_HANDLE;
VkImageView vk_image_view_ = VK_NULL_HANDLE;
VmaAllocation allocation_ = VK_NULL_HANDLE;
public:
VKTexture(const char *name) : Texture(name)
{
}
virtual ~VKTexture() override;
void generate_mipmap() override;
void copy_to(Texture *tex) override;
@ -34,10 +42,36 @@ class VKTexture : public Texture {
/* TODO(fclem): Legacy. Should be removed at some point. */
uint gl_bindcode_get() const override;
void image_bind(int location);
VkImage vk_image_handle() const
{
return vk_image_;
}
VkImageView vk_image_view_handle() const
{
return vk_image_view_;
}
protected:
bool init_internal() override;
bool init_internal(GPUVertBuf *vbo) override;
bool init_internal(const GPUTexture *src, int mip_offset, int layer_offset) override;
private:
/** Is this texture already allocated on device.*/
bool is_allocated();
/**
* Allocate the texture of the device. Result is `true` when texture is successfully allocated
* on the device.
*/
bool allocate();
VkImageViewType vk_image_view_type() const;
};
static inline VKTexture *unwrap(Texture *tex)
{
return static_cast<VKTexture *>(tex);
}
} // namespace blender::gpu

34
source/blender/gpu/vulkan/vk_vertex_buffer.cc
View File

@ -7,6 +7,8 @@
#include "MEM_guardedalloc.h"
#include "vk_shader.hh"
#include "vk_shader_interface.hh"
#include "vk_vertex_buffer.hh"
namespace blender::gpu {
@ -16,8 +18,18 @@ VKVertexBuffer::~VKVertexBuffer()
release_data();
}
void VKVertexBuffer::bind_as_ssbo(uint /*binding*/)
void VKVertexBuffer::bind_as_ssbo(uint binding)
{
VKContext &context = *VKContext::get();
if (!buffer_.is_allocated()) {
allocate(context);
}
VKShader *shader = static_cast<VKShader *>(context.shader);
const VKShaderInterface &shader_interface = shader->interface_get();
const ShaderInput *shader_input = shader_interface.shader_input_get(
shader::ShaderCreateInfo::Resource::BindType::STORAGE_BUFFER, binding);
shader->pipeline_get().descriptor_set_get().bind_as_ssbo(*this, shader_input);
}
void VKVertexBuffer::bind_as_texture(uint /*binding*/)
@ -32,8 +44,17 @@ void VKVertexBuffer::update_sub(uint /*start*/, uint /*len*/, const void * /*dat
{
}
void VKVertexBuffer::read(void * /*data*/) const
void VKVertexBuffer::read(void *data) const
{
VKContext &context = *VKContext::get();
VKCommandBuffer &command_buffer = context.command_buffer_get();
command_buffer.submit();
void *mapped_memory;
if (buffer_.map(context, &mapped_memory)) {
memcpy(data, mapped_memory, size_used_get());
buffer_.unmap(context);
}
}
void VKVertexBuffer::acquire_data()
@ -64,4 +85,13 @@ void VKVertexBuffer::duplicate_data(VertBuf * /*dst*/)
{
}
void VKVertexBuffer::allocate(VKContext &context)
{
buffer_.create(context,
size_used_get(),
usage_,
static_cast<VkBufferUsageFlagBits>(VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
VK_BUFFER_USAGE_VERTEX_BUFFER_BIT));
}
} // namespace blender::gpu

12
source/blender/gpu/vulkan/vk_vertex_buffer.hh
View File

@ -9,9 +9,13 @@
#include "gpu_vertex_buffer_private.hh"
#include "vk_buffer.hh"
namespace blender::gpu {
class VKVertexBuffer : public VertBuf {
VKBuffer buffer_;
public:
~VKVertexBuffer();
@ -22,12 +26,20 @@ class VKVertexBuffer : public VertBuf {
void update_sub(uint start, uint len, const void *data) override;
void read(void *data) const override;
VkBuffer vk_handle() const
{
return buffer_.vk_handle();
}
protected:
void acquire_data() override;
void resize_data() override;
void release_data() override;
void upload_data() override;
void duplicate_data(VertBuf *dst) override;
private:
void allocate(VKContext &context);
};
} // namespace blender::gpu