High Performance Computing Education: Current Challenges …
https://dl.acm.org/doi/pdf/10.1145/3437800.3439203
Rajendra K. Raj∗Carol J. Romanowski†John Impagliazzo†Sherif G. AlyBrett A. BeckerJuan ChenSheikh GhafoorNasser GiacamanCruz IzuShahram RahimiMichael P. RobsonNeena ThotaABSTRACTCCS CONCEPTS2.2 Contrasting HPC and PDC2.4.1 What To Teach? HPC plays an important role in a wide range of application fields, some within its own CS discipline, such as artificial intelligence and the Internet of things, and others across disciplines such as data science and bioinformatics. The diversity of domains makes teaching HPC dificult [90]. Experts in one domain lack necessary expertise in other HPC domains [76], which will greatly limit students’ competency and versatility in the broader HPC field. Breaking the barriers between these various knowledge fields is crucial to produce multi-skilled talented graduates who can span multiple application fields. But developing and carrying out a uniform HPC curricular system is not feasible; we would not expect to build a complete HPC curricula across all possible fields. Competency-based learning can bridge the gap among diferent fields and is an efective way to mitigate this problem.2.4.3 Who Teaches and Who Learns? HPC remains a challenge for both computing students and instructors.3.1 Competency Areas3.3 Defining Competencies for HPC3.3.1 HPC Competency Structure. Generating a list of HPC com-petencies follows the same structure as previously mentioned for regular computing areas. Skills, dispositions, and tasks remain the same; the only component that changes is knowledge. Therefore, we will refer to the most recent knowledge areas developed by the Computing Curriculum Development and Educational Resources (CDER) center at Georgia State University [25]. Table 2 provides a summary of these knowledge elements related to HPC. The ele-ments reflect the experience of the authors and those attributed to the CDER project.3.3.2 Sample Basic Competencies. The following sample compe-tencies apply to both non-computing and computing majors.3.3.3 Sample Competencies for Computing and STEM Majors. The following sample competencies are intended for computing and STEM majors.4.1 Educational Resources4.2 Infrastructure Resources for HPCEd4.2.1 Hardware. The overwhelming majority of modern devices, including smart phones, include multiple processing cores. There-fore, beginners can use standard resources such as laptops, desktop PCs, and general purpose lab computers.4.2.2 Software. In addition to hardware resources, a mix of soft-ware infrastructure is required to support basic parallel education. Parallelism exists or can be supported in a variety of modern languages from Java to Python. Java supports a native threading abstraction as well as several libraries which can be used for dis-tributed computing, e.g., MPI support. Python also supports a vari-ety of high performance libraries, both at the shared and distributed memory levels. Julia is another useful parallel programming lan-guage that support the three main features for concurrent and parallel programming: asynchronous tasks, multi-threading, and distributed computing [73].4.3 Overview of Educational Papers5 HPC OPPORTUNITIES IN COMPUTING5.4 Other Computing Areas6 HPC OPPORTUNITIES IN OTHER DOMAINS7 RECOMMENDATIONS7.1 Building a Professional HPC Educator Community7.2 Developing an Understanding of HPC7.3 Building HPC Competencies7.4 Building Infrastructures for HPC Education7.5 Providing HPC Experiences in Contemporary ComputingIncreasing interdisciplinary knowledge and application ability is a very important educational goal but is dificult to achieve in the typical crowded computing curricula. In fact, most computing majors do not have enough cross-domain knowledge, which limits their development in the HPC field [61]. Interdisciplinary area for non-computing majors. Co...See more on dl.acm.org Increasing interdisciplinary knowledge and application ability is a very important educational goal but is dificult to achieve in the typical crowded computing curricula. In fact, most computing majors do not have enough cross-domain knowledge, which limits their development in the HPC field [61]. Interdisciplinary area for non-computing majors. Co... Author: Rajendra K. Raj, Carol J. Romanowski, John Impagliazzo, Sherif G. Aly, Brett A. Becker, Juan Chen, S... Publish Year: 2020
Increasing interdisciplinary knowledge and application ability is a very important educational goal but is dificult to achieve in the typical crowded computing curricula. In fact, most computing majors do not have enough cross-domain knowledge, which limits their development in the HPC field [61]. Interdisciplinary area for non-computing majors. Co...
Author: Rajendra K. Raj, Carol J. Romanowski, John Impagliazzo, Sherif G. Aly, Brett A. Becker, Juan Chen, S...
Publish Year: 2020
DA: 19 PA: 4 MOZ Rank: 93