Book Image

Hands-On GPU Computing with Python

By : Avimanyu Bandyopadhyay
Book Image

Hands-On GPU Computing with Python

By: Avimanyu Bandyopadhyay

Overview of this book

GPUs are proving to be excellent general purpose-parallel computing solutions for high-performance tasks such as deep learning and scientific computing. This book will be your guide to getting started with GPU computing. It begins by introducing GPU computing and explaining the GPU architecture and programming models. You will learn, by example, how to perform GPU programming with Python, and look at using integrations such as PyCUDA, PyOpenCL, CuPy, and Numba with Anaconda for various tasks such as machine learning and data mining. In addition to this, you will get to grips with GPU workflows, management, and deployment using modern containerization solutions. Toward the end of the book, you will get familiar with the principles of distributed computing for training machine learning models and enhancing efficiency and performance. By the end of this book, you will be able to set up a GPU ecosystem for running complex applications and data models that demand great processing capabilities, and be able to efficiently manage memory to compute your application effectively and quickly.

Related Content you might be interested in

Current Title:

Small book image
Hands-On GPU Computing with Python
Avimanyu Bandyopadhyay
May, 2019 | 452 pages
Small book image
Hands-On GPU Programming with Python and CUDA
Dr Brian Tuomanen
Nov, 2018 | 310 pages
Small book image
CUDA Cookbook
Bharatkumar Sharma | Jack Han
Sep, 2019 | 508 pages
Small book image
Hands-On GPU-Accelerated Computer Vision with OpenCV and CUDA
Bhaumik Vaidya
Sep, 2018 | 380 pages
Table of Contents (17 chapters)
Preface
Preface
Who this book is for
What this book covers
To get the most out of this book
Get in touch
Free Chapter
1
Section 1: Computing with GPUs Introduction, Fundamental Concepts, and Hardware
Section 1: Computing with GPUs Introduction, Fundamental Concepts, and Hardware
2
Introducing GPU Computing
Introducing GPU Computing
The world of GPU computing beyond PC gaming
Conventional CPU computing – before the advent of GPUs
How the gaming industry made GPU computing affordable for individuals
The emergence of full-fledged GPU computing
The simplicity of Python code and the power of GPUs – a dual advantage
How GPUs empower science and AI in current times
The social impact of GPUs
Summary
Further reading
3
Designing a GPU Computing Strategy
Designing a GPU Computing Strategy
Getting started with the hardware
Building your first GPU-enabled parallel computer – minimum system requirements
Liquid cooling – should you consider it?
Branded GPU-enabled PCs
Why not DIY?
Entry-level budget
Mid-range budget
High-end budget
Summary
Further reading
4
Setting Up a GPU Computing Platform with NVIDIA and AMD
Setting Up a GPU Computing Platform with NVIDIA and AMD
GPU manufacturers
Computing on NVIDIA GPUs
Computing on AMD APUs and GPUs
Comparing GPU programmable platforms on NVIDIA and AMD
Current models from both brands that are ideal for GPU computing
An enthusiast's guide to GPU computing hardware
Summary
Further reading
5
Section 2: Hands-On Development with GPU Programming
Section 2: Hands-On Development with GPU Programming
6
Fundamentals of GPU Programming
Fundamentals of GPU Programming
GPU-programmable platforms
Basic CUDA concepts
Basic ROCm concepts
The Anaconda Python distribution for package management and deployment
GPU-enabled Python programming
Summary
Further reading
7
Setting Up Your Environment for GPU Programming
Setting Up Your Environment for GPU Programming
Choosing a suitable IDE for your Python code
PyCharm – an IDE exclusively made for Python
Installing PyCharm
Alternative IDEs for Python – PyDev and Jupyter
Installing the PyDev Python IDE for Eclipse
Installing Jupyter Notebook and Jupyter Lab
Summary
Further reading
8
Working with CUDA and PyCUDA
Working with CUDA and PyCUDA
Technical requirements
Understanding how CUDA-C/C++ works via a simple example
Installing PyCUDA for Python within an existing CUDA environment
Configuring PyCUDA on your Python IDE
How computing in PyCUDA works on Python
Comparing PyCUDA to CUDA – an introductory perspective on reduction
Writing your first PyCUDA programs to compute a general-purpose solution
Useful exercise on computational problem solving
Summary
Further reading
9
Working with ROCm and PyOpenCL
Working with ROCm and PyOpenCL
Technical requirements
Understanding how ROCm-C/C++ works with hipify, HIP, and OpenCL
Installing PyOpenCL for Python (AMD and NVIDIA)
Configuring PyOpenCL on your Python IDE
How computing in PyOpenCL works on Python
Comparing PyOpenCL to HIP and OpenCL – revisiting the reduction perspective
Writing your first PyOpenCL programs to compute a general-purpose solution
Useful exercise on computational problem solving
Summary
Further reading
10
Working with Anaconda, CuPy, and Numba for GPUs
Working with Anaconda, CuPy, and Numba for GPUs
Technical requirements
Understanding how Anaconda works with CuPy and Numba
Installing CuPy and Numba for Python within an existing Anaconda environment
Configuring CuPy on your Python IDE
How computing in CuPy works on Python
Configuring Numba on your Python IDE
How computing in Numba works on Python
Writing your first CuPy and Numba enabled accelerated programs to compute GPGPU solutions
Interoperability between CuPy and Numba within a single Python program
Comparing CuPy to NumPy and CUDA
Comparing Numba to NumPy, ROCm, and CUDA
Useful exercise on computational problem solving
Summary
Further reading
11
Section 3: Containerization and Machine Learning with GPU-Powered Python
Section 3: Containerization and Machine Learning with GPU-Powered Python
12
Containerization on GPU-Enabled Platforms
Containerization on GPU-Enabled Platforms
Programmable environments
System-wide (open) environments
Virtual (closed) environments
Virtualization
Local containers
Cloud containers
Summary
Further reading
13
Accelerated Machine Learning on GPUs
Accelerated Machine Learning on GPUs
Technical requirements
The significance of Python in AI – the dual advantage
Exploring machine learning training modules
Introducing machine learning frameworks
Introducing TensorFlow
Introducing PyTorch
Installing TensorFlow and PyTorch for GPUs
Configuring TensorFlow on PyCharm and Google Colab
Configuring PyTorch on PyCharm and Google Colab
Machine learning with TensorFlow and PyTorch
Writing your first GPU-accelerated machine learning programs
Revisiting our computational exercises with a machine learning approach
Summary
Further reading
14
GPU Acceleration for Scientific Applications Using DeepChem
GPU Acceleration for Scientific Applications Using DeepChem
Technical requirements
Decoding scientific concepts for DeepChem
Multiple ways to install DeepChem
Configuring DeepChem on PyCharm
Testing an example from the DeepChem repository
Developing your own deep learning framework like DeepChem – a brief outlook
Summary
Final thoughts
References
15
Other Books You May Enjoy
Other Books You May Enjoy
Leave a review - let other readers know what you think
Appendix A
Appendix A
GPU-accelerated machine learning in Python – benchmark research
GPU-accelerated machine learning with Python applied to cancer research
Deep Learning with GPU-accelerated Python for applied computer vision – Pavement Distress

Understanding how CUDA-C/C++ works via a simple example

By now, you must be aware of the computational advantages of CUDA C/C++ as per our earlier discussions. C/C++ coupled with CUDA allows you to modify parts of your source code to accelerate your computational results. The primary steps necessary for implementing CUDA code will be explored through a GPU program.

Please manually type in the code used in this book on your IDE from this point onward. Directly copying and pasting from the PDF will ruin the indentations in the code and make it unready to deploy.

First, let's look into the following conventional C++ program that multiplies two array elements using double precision. We'll run the kernel on 500 million elements on the CPU. All the elements of the p and q arrays are set to 24 and 12 respectively.

The following is the C++ program we've just described ...

Previous Section
End of Section 3
Next Section