Louisville, Kentucky
June 20, 2010
June 20, 2010
June 23, 2010
2153-5965
Electrical and Computer
12
15.983.1 - 15.983.12
10.18260/1-2--15945
https://peer.asee.org/15945
582
Problem-Base Learning of Multicore Programming Department of Electrical and Computer Engineering Southern Illinois University Carbondale Carbondale, IL 62901
1. Introduction
The computer industry is rapidly moving towards the multicore processors. Manycore processors have been widely used in all computing domains, including desktop, server and embedded computing systems. A multicore processor combines two or more independent processors into a single package, which is capable of executing multiple threads simultaneously. The L2 cache on a multicore processor can be either private or shared, as depicted in Figure 1 (a) and (b), respectively. Clearly, multicore processors can naturally benefit multithreaded programs by running them on different cores concurrently to improve the throughput. However, unlike other advances of microprocessors aiming at the transparent increase of single-threaded performance (e.g., frequency scaling, pipelines, caches, and superscalar architectures), multicore processors cannot automatically reduce the latency of single-threaded programs. In many cases, there is no way to effectively utilize the performance of additional processor cores or hardware threads without writing explicitly multithreaded programs [1]. In other words, the advent of multicore processors necessitates a fundamental shift of software development paradigm, from the traditional single-threaded programming to multithreaded programming, which is a great challenge.
Thread Thread 2 Thread Thread 2 1 1
CPU CPU CPU CPU and L1 $ and L1 $ and L1 $ and L1 $
L2 L2 Shared L2 Cache
System System L3$/Memory Bus L3$/Memory Bus (a) (b)
Figure 1. A dual-core processor with (a) private L2 caches or (b) shared L2 cache. To prepare students for the coming era of multicore computing, I propose to develop a multicore programming course for our senior and graduate students. The goals of this project are to: • Create a course on multicore programming for general-purpose applications, which can be potentially adapted or adopted by the computer engineering or computer science departments at other institutions. • Investigate the effective use of problem-based learning (PBL) in this course to overcome the difficulty of teaching and learning MTP. More specifically, this project is expected to provide useful insights into several key PBL problems,
Zhang, W. (2010, June), Problem Base Learning Of Multi Core Programming Paper presented at 2010 Annual Conference & Exposition, Louisville, Kentucky. 10.18260/1-2--15945
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2010 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015