Encyclopedia of Parallel Computing Surveys Concepts Behind Shift to Parallelism

7/25/2016

Prof. Padua led the effort to create a 4-volume series surveying the concepts behind shift to parallel computing.

Written by

University of Illinois computer science professor David Padua served as Editor-in-Chief for a new encyclopedic effort to survey the concepts behind the significant shift towards parallel computing in today’s computer industry. The 4 volume Encyclopedia of Parallel Computing contains more than 300 entries on topics related to the critical demand for continued advances in parallel programming. Dozens of contributors from all over the world were integral to the success of the project, including leading international experts in the field, like Illinois computer science professors William Gropp and Marc Snir, professor emeritus David J. Kuck, as well as Jack Dongarra, Michael Flynn, David E. Shaw, Guy L. Steele Jr., among others.  Editorial board members included Illinois professors Sarita Adve, Maria Garzaran, William Gropp, and Laxmikant Kale, and a total of 19 Illinois faculty, staff, and students contributed to entries in the Encyclopedia.

David Padua
David Padua

 

“Parallel computing has already impacted or will soon impact everyone who uses a computing device, from supercomputers to laptops, tablets, and smart phones,” said Padua.

Today’s supercomputers are massively parallel machines with thousands of processors. The fastest supercomputer today uses 705,024 processing cores, capable of 10.51 quadrillion calculations per second. Ten years ago the world’s fastest supercomputer used a total of 8,192 processing cores and was only capable of 12.3 trillion calculations per second, almost one thousand times less powerful. This type of accelerated parallelism is critical to science and engineering, enabling discoveries and designs that would not be possible otherwise. For consumer and mobile devices, parallelism is the only viable strategy for continued performance gains, while also allowing chipmakers to optimize for energy efficiency.

With the need for parallelism at an all-time high, the Encyclopedia of Parallel Computing provides researchers and developers with an authoritative reference that pulls together the tools necessary to take advantage of these pioneering concepts.

"This monumental work will be an invaluable resource for practitioners and students in all areas of computer science,” said Alex Nicolau of UC Irvine. “In today's world where parallel computing is ubiquitous--from desktops to cell phones and game consoles--this reference work will be indispensable."

Key concepts in the Encyclopedia for professionals, researchers and students of Parallel Computing include:

  • Programming models and programming languages
  •  
  • Debugging and race detection
  • Laws of parallel computing
  •  
  • Theoretical models of computation
  •  
  • Supercomputer/high-performance computing machines
  •  
  • Interconnection networks

 


Share this story

This story was published July 25, 2016.