Book Image

Learning Vulkan

By : Parminder Singh
Book Image

Learning Vulkan

By: Parminder Singh

Overview of this book

Vulkan, the next generation graphics and compute API, is the latest offering by Khronos. This API is the successor of OpenGL and unlike OpenGL, it offers great flexibility and high performance capabilities to control modern GPU devices. With this book, you'll get great insights into the workings of Vulkan and how you can make stunning graphics run with minimum hardware requirements. We begin with a brief introduction to the Vulkan system and show you its distinct features with the successor to the OpenGL API. First, you will see how to establish a connection with hardware devices to query the available queues, memory types, and capabilities offered. Vulkan is verbose, so before diving deep into programing, you’ll get to grips with debugging techniques so even first-timers can overcome error traps using Vulkan’s layer and extension features. You’ll get a grip on command buffers and acquire the knowledge to record various operation commands into command buffer and submit it to a proper queue for GPU processing. We’ll take a detailed look at memory management and demonstrate the use of buffer and image resources to create drawing textures and image views for the presentation engine and vertex buffers to store geometry information. You'll get a brief overview of SPIR-V, the new way to manage shaders, and you'll define the drawing operations as a single unit of work in the Render pass with the help of attachments and subpasses. You'll also create frame buffers and build a solid graphics pipeline, as well as making use of the synchronizing mechanism to manage GPU and CPU hand-shaking. By the end, you’ll know everything you need to know to get your hands dirty with the coolest Graphics API on the block.
Table of Contents (18 chapters)
Learning Vulkan
Credits
About the Author
Acknowledgments
About the Reviewer
www.PacktPub.com
Preface

Understanding compute pipelines


A compute pipeline consists of a single static compute shader stage and the pipeline layout. The compute shader stage is capable of doing massive parallel arbitrary computations. On the other hand, the pipeline layout connects the compute pipeline to the descriptor using the layout bindings. The vkCreateComputePipeline() API can be used to create a compute pipeline:

VkResult vkCreateComputePipelines( 
   VkDevice                            device, 
   VkPipelineCache                     pipelineCache, 
   uint32_t                            createInfoCount, 
   const VkComputePipelineCreateInfo*  pCreateInfos, 
   const VkAllocationCallbacks*        pAllocator, 
   VkPipeline*                         pPipelines); 

The following table describes the various fields of the vkCreateGraphicsPipelines() API:

Parameters

Description

device

This is the logical device (VkDevice) from which the compute pipeline object is going to...