Parallel programming with pictures is a Snap

For decades, computing speeds seemingly doubled every 24 months by increasing the processor clock speed, thus giving software a “free ride” to better performance. This free ride, however, effectively ended by the mid-2000s. With clock speeds having plateaued and computational horsepower instead incr...

Full description

Saved in:
Bibliographic Details
Published in:Journal of parallel and distributed computing Vol. 105; pp. 150 - 162
Main Authors: Feng, Annette, Gardner, Mark, Feng, Wu-chun
Format: Journal Article
Language:English
Published: Elsevier Inc 01.07.2017
Subjects:
Block-based programming
Computer science education
Explicit parallel computing
Languages for PDC and HPC
Parallel computational patterns
Pedagogical tools
Programming environments
Visual programming
Visual programming
Explicit parallel computing
Pedagogical tools
Computer science education
Block-based programming
Parallel computational patterns
Languages for PDC and HPC
Programming environments
ISSN:0743-7315, 1096-0848
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:For decades, computing speeds seemingly doubled every 24 months by increasing the processor clock speed, thus giving software a “free ride” to better performance. This free ride, however, effectively ended by the mid-2000s. With clock speeds having plateaued and computational horsepower instead increasing due to increasing the number of cores per processor, the vision for parallel computing, which started more than 40 years ago, is a revolution that has now (ubiquitously) arrived. In addition to traditional supercomputing clusters, parallel computing with multiple cores can be found in desktops, laptops, and even mobile smartphones. This ubiquitous parallelism in hardware presents a major challenge: the difficulty in easily extracting parallel performance via current software abstractions. Consequently, this paper presents an approach that reduces the learning curve to parallel programming by introducing such concepts into a visual (but currently sequential) programming language called Snap!, which was inspired by MIT’s Scratch project. Furthermore, our proposed visual abstractions can automatically generate parallel code for the end user to run in parallel on a variety of platforms from personal computing devices to supercomputers. Ultimately, this work seeks to increase parallel programming literacy so that users, whether novice or experienced, may leverage a world of ubiquitous parallelism to enhance productivity in all walks of life, including the sciences, engineering, commerce, and liberal arts. •Introduction of explicit parallel-programming constructs into a picture-programming environment.•Block-based programming abstractions to reduce the learning curve to parallel programming.•Automated generation of OpenMP code from our parallel picture-programming environment.
ISSN:0743-7315
1096-0848
DOI:10.1016/j.jpdc.2017.01.018