Exploring concurrency using the parallel analysis tool
Proceedings of the 43rd ACM technical symposium on Computer Science Education
Integrating data-intensive cloud computing with multicores and clusters in an HPC course
Proceedings of the 17th ACM annual conference on Innovation and technology in computer science education
| Hi-index | 0.00 |
Educators in Computer Science (CS) generally agree that teaching concurrency can be difficult. CS programs typically offer parallel and distributed computing topics as advanced courses. A potential alternative approach is to provide instruction on parallelism early in the undergraduate curriculum, emphasizing conceptual design rather than implementation issues. This introduction of "parallel-thinking" to beginning CS undergraduates represents an innovation and significant extension to existing standard Computer Science curricula. The research described in this paper investigated the feasibility of integrating parallel computing concepts into a first-year CS course. To quantitatively assess student comprehension of parallel computing, an experimental two-factor mixed group design educational study was conducted to evaluate a control group and two instructional interventions: (1) lecture only, and (2) lecture with laboratory work using a software visualization Parallel Analysis Tool (PAT) specifically designed for this project. The Perceptions of Parallelism Survey (PoPS), a new evaluation instrument developed for this study and modeled after the Force Concept Inventory (FCI), was used to measure student learning. The results from this educational study show a statistically significant main effect among the repeated measures. PoPS results measured during the ninth week of the course reveal that performance levels remained high compared to pre-course performance scores.