Hwu co-authors ground-breaking textbook on parallel programming
Laurie Talkington, Coordinated Science Lab
- ECE Prof. Wen-mei Hwu and David Kirk of NVIDIA have written a new text on parallel processing.
- The book will give programmers the tools to optimize GPU.
- The text developed out of a course taught by Hwu and Kirk in 2007 at Illinois.
ECE Professor Wen-mei W. Hwu has co-authored what many in the industry believe will be an important new textbook that breaks down the complexities of parallel programming and the graphics processing unit (GPU) architecture to enable programmers to address the critical challenges of massive parallelism.
Programming Massively Parallel Processors: A Hands-on Approach, was co-authored by Hwu, principal investigator of Illinois’s CUDA Center of Excellence, and David Kirk, an NVIDIA Fellow and former Chief Scientist.
“The reason we wrote that book is because the semiconductor industry is moving in a big way into parallel computing,” said Hwu in a video on the publisher’s Web site. He noted that the industry road map expects that by 2016, parallel processing will be able to run up to 6,000 times faster than current sequential code.
Hwu said that the book was written to give programmers at all levels and across all scientific disciplines the ability to “think parallel” and optimize the GPU.
“This book is written, not to just educate computer science and computer engineering students, but it’s really designed to give students in all [business, science, and engineering] disciplines the ability to think parallel, to be able to use these techniques in their own work,” said Hwu.
The text was developed from one of the nation’s first university courses to offer a comprehensive approach to massively parallel programming. Originally co-taught by the authors at Illinois in 2007, the course materials have since been tested and refined for presentation in venues and media around the world. Various online modules have been adopted into the curricula at over 100 universities around the world.
The book is one of the first major publications by the CUDA Centers of Excellence worldwide, which are dedicated to providing global access to education in massively parallel processing. While Illinois hosts the original CUDA Center, others are hosted at the universities of Utah, Harvard, and Tennessee in the USA; the University of Cambridge in the UK; National Taiwan University; and in China, Tsinghua University and the Chinese Academy of Sciences.
More information on Programming Massively Parallel Processors: A Hands-on Approach can be found at the publisher’s Web site , including sample chapters and ancillary teaching materials.
Hwu is the W. J. "Jerry" Sanders III-Advanced Micro Devices Endowed Chair in Electrical & Computer Engineering. He is a researcher in the Coordinated Science Laboratory. His IMPACT Research Group is well known for development of the IMPACT Compiler and other novel computing technologies now widely used in industry and academic research. He is a leader of many parallel computing research and teaching initiatives at the University of Illinois, which are aligned through Parallel@Illinois. He is director of the world's first NVIDIA CUDA Center of Excellence, co-director of the Intel-Microsoft-funded Universal Parallel Computing Research Center (UPCRC), and co-principal investigator for hardware of the NSF Petascale Computer Project, “Blue Waters” in collaboration with IBM. He is a Fellow of the IEEE and ACM and author of over 200 peer-reviewed articles.
Kirk is a Fellow and former Chief Scientist of NVIDIA. He led the development of graphics technology for many popular consumer entertainment platforms and holds over 50 patents and patent applications relating to graphics design. He has published more than 50 articles on graphics technology and was elected to the National Academy of Engineering in 2006. In 2002, he was awarded the SIGGRAPH Computer Graphics Achievement Award for his role in bringing high-performance computer graphics systems to the mass market.