Quantitative performance evaluation of high-end computing platforms

We have a wide variety of systems and configurations to choose for our computing needs; however, there is no knowledge base of which systems have better performance for a given business, scientific or an engineering application. In this paper, we benchmarked the performance of commonly available hig...

Full description

Saved in:
Bibliographic Details
Main Authors: Veeraraghavan, K., Sharif, H., Nicoll, A., Ali, H.
Format: Conference Proceeding
Language:English
Subjects:
Benchmark testing
Computational modeling
Computer architecture
Educational institutions
High performance computing
Information science
Knowledge engineering
Operating systems
Performance evaluation
System testing
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We have a wide variety of systems and configurations to choose for our computing needs; however, there is no knowledge base of which systems have better performance for a given business, scientific or an engineering application. In this paper, we benchmarked the performance of commonly available high-end systems using industry standard benchmark suites. Since most of the currently available benchmarks do not provide a complete evaluation of all system components, we chose to use different benchmarks to measure the performance of low-level, kernel and application specific areas. We have described the architecture and configuration of each system tested and also presented the benchmark results, which reveal the best suited platform for application requirements
ISSN:2154-0357
2154-0373
DOI:10.1109/EIT.2005.1626984