Abstract
For many years, C has been known as a fast, yet unfriendly language. Similarly, Java presents its own trade-offs, including more advanced language features at the cost of slower execution. As Java implementations continue to mature, this distinction has become less clear. Special hardware, better libraries and more sophisticated compilers have placed Java-based program execution times in the same realm as C-based programs. This paper demonstrates that superior performance is obtained by carefully selecting the appropriate language for implementing the system. In some cases, this will involve both languages interacting synergistically. For the SciMark 2.0 benchmark suite, using the Java Native Interface (JNI) increases performance by 2.29X for the best case and 1.29X on average compared to the C implementation on a Sharp Zaurus PDA. Simultaneously, the energy improvement for the best case is 2.26X and 1.24X on average.
Original language | English (US) |
---|---|
Title of host publication | Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the INternational Conference on VLSI, VLSI'04 |
Editors | H.R. Arabnia, M. Guo, L.T. Yang |
Pages | 185-191 |
Number of pages | 7 |
State | Published - Dec 1 2004 |
Event | Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the International Conference on VLSI, VLSI'04 - Las Vegas, NV, United States Duration: Jun 21 2004 → Jun 24 2004 |
Other
Other | Proceedings of the International Conference on Embedded Systems and Applications ESA'04 - Proceedings of the International Conference on VLSI, VLSI'04 |
---|---|
Country/Territory | United States |
City | Las Vegas, NV |
Period | 6/21/04 → 6/24/04 |
All Science Journal Classification (ASJC) codes
- Engineering(all)