Introduction to Concurrent Programming with GPUs

Overview

This course will help prepare students for developing code that can process large amounts of data in parallel. It will focus on foundational aspects of concurrent programming, such as CPU/GPU architectures, multithreaded programming in C and Python, and an introduction to CUDA software/hardware.

Syllabus

Course Overview

The purpose of this foundational module is to prepare students for success within the GPU Programming Specialization. To this end, the module presents the course's overall goals and technical expectations and guides utilizing the Coursera lab environment. Students will conclude by completing an introductory C++ programming assignment to ensure they are equipped for the curriculum.

Core Principles of Parallel Programming on CPUs and GPUs

In order to create software that process greater amounts of data at faster speeds, software operating systems, programming languages, and frameworks require strategies for accessing and modification of data in a manner that maximizes speed, while minimizing the possibility of data being in incorrect states. In this module, students will be presented canonical concurrency problems such as the Dining Philosophers. Additionally, they will learn how operating systems and programming languages handle these problems, and discuss real world big data concurrency applications.

Introduction to Parallel Programming with C and Python

Modern programming languages allow developers to create software with complex logic for manipulation of data in parallel, taking advantage of the multiple CPU cores in most computers. Students will develop simple software, written in the C++ and Python 3 programming languages, that process data sets concurrently.

NVidia GPU Hardware/Software

In order to write effective GPU-accelerated software, a developer must first understand the underlying hardware architecture and the software ecosystem that enables it. In this module, students will be presented with the key hardware architectures of NVidia GPUs. Additionally, they will learn about the CUDA software layers, the installation and compilation process, and the core differences between the Runtime and Driver APIs.

Introduction to GPU Programming

The purpose of this module is to transition from theory to practice by introducing the fundamentals of GPU programming. Students will learn the essential CUDA software keywords and syntax required for writing device code. They will also be guided through best practices for structuring both simple and complex projects, applying these concepts in hands-on assignments within common development environments.