Building a Parallel Computer Using Cheap PCs

From Simple Sci Wiki
Revision as of 02:48, 24 December 2023 by SatoshiNakamoto (talk | contribs) (Created page with "Title: Building a Parallel Computer Using Cheap PCs Abstract: This research explores the possibility of creating a powerful parallel computer using multiple cheap personal computers (PCs). The study discusses the configuration, performance, and applications of a PC cluster built for physics research. Main Research Question: Can a parallel computer be built using multiple cheap PCs that can compete with traditional supercomputers? Methodology: The researchers built a P...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Title: Building a Parallel Computer Using Cheap PCs

Abstract: This research explores the possibility of creating a powerful parallel computer using multiple cheap personal computers (PCs). The study discusses the configuration, performance, and applications of a PC cluster built for physics research.

Main Research Question: Can a parallel computer be built using multiple cheap PCs that can compete with traditional supercomputers?

Methodology: The researchers built a PC cluster with ten PCs, each equipped with two 500 MHz Pentium III processors. They used a fast ethernet switch to handle inter-node communication. The cluster ran on the Linux operating system, which allowed for multiple processors and network file sharing. They used the Message Passing Interface (MPI) for parallel processing.

Results: The researchers found that their PC cluster could perform parallel computations as effectively as traditional supercomputers. They also noted that the cluster's cost was significantly lower than that of supercomputers, making it an attractive option for academic research budgets.

Implications: This research suggests that building a parallel computer using cheap PCs is a viable and cost-effective solution for organizations with limited budgets. It also demonstrates that the performance of such a cluster can compete with that of traditional supercomputers, making it a promising option for various fields of research and commercial applications.

Link to Article: https://arxiv.org/abs/0109004v1 Authors: arXiv ID: 0109004v1