Analysis of Sequential and Parallel Execution of Simpson's Rule for Numerical Integration in Java
Marko M. Smilic, Dejan Milić, Caslav M. Stefanovic, Sanja Đurović, Danijel Đošić, Marjan M. Matejić · 2025
This paper analyses the sequential and parallel execution of Simpson's Rule in Java for numerical integration, focusing on time analysis, speedup, efficiency, and overhead. Using multi-core processors, we examine performance improvements for computing the Fresnel integral. The implementation is evaluated based on execution time, hardware configuration, software setup, and accuracy relative to the threshold value. Results show that parallel execution significantly reduces computation time and improves optimally configured accuracy. We also discuss the trade-offs of parallelism, such as overhead and diminishing returns, and its implications for high-performance computing in Java..