Vulkan Cookbook

By : Pawel Lapinski

Vulkan Cookbook

By: Pawel Lapinski

Overview of this book

Vulkan is the next generation graphics API released by the Khronos group. It is expected to be the successor to OpenGL and OpenGL ES, which it shares some similarities with such as its cross-platform capabilities, programmed pipeline stages, or nomenclature. Vulkan is a low-level API that gives developers much more control over the hardware, but also adds new responsibilities such as explicit memory and resources management. With it, though, Vulkan is expected to be much faster. This book is your guide to understanding Vulkan through a series of recipes. We start off by teaching you how to create instances in Vulkan and choose the device on which operations will be performed. You will then explore more complex topics such as command buffers, resources and memory management, pipelines, GLSL shaders, render passes, and more. Gradually, the book moves on to teach you advanced rendering techniques, how to draw 3D scenes, and how to improve the performance of your applications. By the end of the book, you will be familiar with the latest advanced techniques implemented with the Vulkan API, which can be used on a wide range of platforms.

Preface

What this book covers

What you need for this book

Free Chapter

Instance and Devices

Introduction

Downloading Vulkan's SDK

Enabling validation layers

Connecting with a Vulkan Loader library

Preparing for loading Vulkan API functions

Loading functions exported from a Vulkan Loader library

Loading global-level functions

Checking available Instance extensions

Creating a Vulkan Instance

Loading instance-level functions

Enumerating available physical devices

Checking available device extensions

Getting features and properties of a physical device

Checking available queue families and their properties

Selecting the index of a queue family with the desired capabilities

Creating a logical device

Loading device-level functions

Getting a device queue

Creating a logical device with geometry shaders, graphics, and compute queues

Destroying a logical device

Destroying a Vulkan Instance

Releasing a Vulkan Loader library

Image Presentation

Introduction

Creating a Vulkan Instance with WSI extensions enabled

Creating a presentation surface

Selecting a queue family that supports presentation to a given surface

Creating a logical device with WSI extensions enabled

Selecting a desired presentation mode

Getting the capabilities of a presentation surface

Selecting a number of swapchain images

Choosing a size of swapchain images

Selecting desired usage scenarios of swapchain images

Selecting a transformation of swapchain images

Selecting a format of swapchain images

Creating a swapchain

Getting handles of swapchain images

Creating a swapchain with R8G8B8A8 format and a mailbox present mode

Acquiring a swapchain image

Presenting an image

Destroying a swapchain

Destroying a presentation surface

Command Buffers and Synchronization

Introduction

Creating a command pool

Allocating command buffers

Beginning a command buffer recording operation

Ending a command buffer recording operation

Resetting a command buffer

Resetting a command pool

Submitting command buffers to a queue

Synchronizing two command buffers

Checking if processing of a submitted command buffer has finished

Waiting until all commands submitted to a queue are finished

Waiting for all submitted commands to be finished

Destroying a fence

Destroying a semaphore

Freeing command buffers

Destroying a command pool

Resources and Memory

Introduction

Creating a buffer

Allocating and binding a memory object for a buffer

Setting a buffer memory barrier

Creating a buffer view

Creating an image

Allocating and binding a memory object to an image

Setting an image memory barrier

Creating an image view

Creating a 2D image and view

Creating a layered 2D image with a CUBEMAP view

Mapping, updating and unmapping host-visible memory

Copying data between buffers

Copying data from a buffer to an image

Copying data from an image to a buffer

Using a staging buffer to update a buffer with a device-local memory bound

Using a staging buffer to update an image with a device-local memory bound

Destroying an image view

Destroying an image

Destroying a buffer view

Freeing a memory object

Destroying a buffer

Descriptor Sets

Introduction

Creating a sampler

Creating a sampled image

Creating a combined image sampler

Creating a storage image

Creating a uniform texel buffer

Creating a storage texel buffer

Creating a uniform buffer

Creating a storage buffer

Creating an input attachment

Creating a descriptor set layout

Creating a descriptor pool

Allocating descriptor sets

Updating descriptor sets

Binding descriptor sets

Creating descriptors with a texture and a uniform buffer

Freeing descriptor sets

Resetting a descriptor pool

Destroying a descriptor pool

Destroying a descriptor set layout

Destroying a sampler

Render Passes and Framebuffers

Introduction

Specifying attachments descriptions

Specifying subpass descriptions

Specifying dependencies between subpasses

Creating a render pass

Creating a framebuffer

Preparing a render pass for geometry rendering and postprocess subpasses

Preparing a render pass and a framebuffer with color and depth attachments

Beginning a render pass

Progressing to the next subpass

Ending a render pass

Destroying a framebuffer

Destroying a render pass

Shaders

Introduction

Converting GLSL shaders to SPIR-V assemblies

Writing vertex shaders

Writing tessellation control shaders

Writing tessellation evaluation shaders

Writing geometry shaders

Writing fragment shaders

Writing compute shaders

Writing a vertex shader that multiplies vertex position by a projection matrix

Using push constants in shaders

Writing texturing vertex and fragment shaders

Displaying polygon normals with a geometry shader

Graphics and Compute Pipelines

Introduction

Creating a shader module

Specifying pipeline shader stages

Specifying a pipeline vertex binding description, attribute description, and input state

Specifying a pipeline input assembly state

Specifying a pipeline tessellation state

Specifying a pipeline viewport and scissor test state

Specifying a pipeline rasterization state

Specifying a pipeline multisample state

Specifying a pipeline depth and stencil state

Specifying a pipeline blend state

Specifying pipeline dynamic states

Creating a pipeline layout

Specifying graphics pipeline creation parameters

Creating a pipeline cache object

Retrieving data from a pipeline cache

Merging multiple pipeline cache objects

Creating a graphics pipeline

Creating a compute pipeline

Binding a pipeline object

Creating a pipeline layout with a combined image sampler, a buffer, and push constant ranges

Creating a graphics pipeline with vertex and fragment shaders, depth test enabled, and with dynamic viewport and scissor tests

Creating multiple graphics pipelines on multiple threads

Destroying a pipeline

Destroying a pipeline cache

Destroying a pipeline layout

Destroying a shader module

Command Recording and Drawing

Introduction

Clearing a color image

Clearing a depth-stencil image

Clearing render pass attachments

Binding vertex buffers

Binding an index buffer

Providing data to shaders through push constants

Setting viewport states dynamically

Setting scissor states dynamically

Setting line width states dynamically

Setting depth bias states dynamically

Setting blend constants states dynamically

Drawing a geometry

Drawing an indexed geometry

Dispatching compute work

Executing a secondary command buffer inside a primary command buffer

Recording a command buffer that draws a geometry with dynamic viewport and scissor states

Recording command buffers on multiple threads

Preparing a single frame of animation

Increasing the performance through increasing the number of separately rendered frames

Helper Recipes

Introduction

Preparing a translation matrix

Preparing a rotation matrix

Preparing a scaling matrix

Preparing a perspective projection matrix

Preparing an orthographic projection matrix

Loading texture data from a file

Loading a 3D model from an OBJ file

Lighting

Introduction

Rendering a geometry with a vertex diffuse lighting

Rendering a geometry with a fragment specular lighting

Rendering a normal mapped geometry

Drawing a reflective and refractive geometry using cubemaps

Adding shadows to the scene

Advanced Rendering Techniques

Introduction

Drawing a skybox

Drawing billboards using geometry shaders

Drawing particles using compute and graphics pipelines

Rendering a tessellated terrain

Rendering a full-screen quad for post-processing

Using input attachments for a color correction post-process effect

Customer Reviews

5 star

4 star

3 star

2 star

1 star

Selecting the index of a queue family with the desired capabilities

Before we can create a logical device, we need to think about what operations we want to perform on it, because this will affect our choice of a queue family (or families) from which we want to request queues.

For simple use cases, a single queue from a family that supports graphics operations should be enough. More advanced scenarios will require graphics and compute operations to be supported, or even an additional transfer queue for very fast memory copying.

In this recipe, we will look at how to search for a queue family that supports the desired type of operations.

How to do it...

Take one of the physical device handles returned by the vkEnumeratePhysicalDevices() function and store it in a variable of type VkPhysicalDevice called physical_device.
Prepare a variable of type uint32_t named queue_family_index. In it, we will store an index of a queue family that supports selected types of operations.
Create a bit field variable of type VkQueueFlags named desired_capabilities. Store the desired types of operations in the desired_capabilities variables--it can be a logical OR operation of any of the VK_QUEUE_GRAPHICS_BIT, VK_QUEUE_COMPUTE_BIT, VK_QUEUE_TRANSFER_BIT or VK_QUEUE_SPARSE_BINDING_BIT values.
Create a variable of type std::vector with VkQueueFamilyProperties elements named queue_families.
Check the number of available queue families and acquire their properties as described in the Checking available queue families and their properties recipe. Store the results of this operation in the queue_families variable.
Loop over all elements of the queue_families vector using a variable of type uint32_t named index.
For each element of the queue_families variable:
1. Check if the number of queues (indicated by the queueCount member) in the current element is greater than zero.
2. Check if the logical AND operation of the desired_capabilities variable and the queueFlags member of the currently iterated element is not equal to zero.
3. If both checks are positive, store the value of an index variable (current loop iteration) in the queue_family_index variable, and finish iterating.
Repeat steps from 7.1 to 7.3 until all elements of the queue_families vector are viewed.

How it works...

First, we acquire the properties of queue families available on a given physical device. This is the operation described in the Checking available queue families and their properties recipe. We store the results of the query in the queue_families variable, which is of std::vector type with VkQueueFamilyProperties elements:

std::vector<VkQueueFamilyProperties> queue_families; 
if( !CheckAvailableQueueFamiliesAndTheirProperties( physical_device, queue_families ) ) { 
  return false; 
}

Next, we start inspecting all elements of a queue_families vector:

for( uint32_t index = 0; index < static_cast<uint32_t>(queue_families.size()); ++index ) { 
  if( (queue_families[index].queueCount > 0) && 
      (queue_families[index].queueFlags & desired_capabilities ) ) { 
    queue_family_index = index; 
    return true; 
  } 
} 
return false;

Each element of the queue_families vector represents a separate queue family. Its queueCount member contains the number of queues available in a given family. The queueFlags member is a bit field, in which each bit represents a different type of operation. If a given bit is set, it means that the corresponding type of operation is supported by the given queue family. We can check for any combination of supported operations, but we may need to search for separate queues for every type of operation. This solely depends on the hardware support and the Vulkan API driver.

To be sure that the data we have acquired is correct, we also check if each family exposes at least one queue.

More advanced real-life scenarios would require us to store the total number of queues exposed in each family. This is because we may want to request more than one queue, but we can't request more queues than are available in a given family. In simple use cases, one queue from a given family is enough.

Vulkan Cookbook

By : Pawel Lapinski

Vulkan Cookbook

By: Pawel Lapinski

Overview of this book

Related Content you might be interested in

Current Title:

Vulkan Cookbook

C++ Game Development By Example

3D Graphics Rendering Cookbook

Mastering Graphics Programming with Vulkan