Comparing Performance of Parallel Implementation of Sorting Algorithms Versus Standard Implementations

Presenter(s)

Allen Martin

Abstract

In this research, sorting algorithms that utilize parallel programming were implemented to determine if parallel programming can be utilized to create an algorithm with a faster execution time. An implementation of the quick sort was developed that utilizes parallel programming via the Fork Join class in Java, and it was compared to the sequential quick sort algorithm. The comparison was made by using arrays of different sizes with unsorted integers. The arrays were sorted through both algorithms and parallel algorithms with various threads. The best algorithm and the best thread count were determined. The efficiency of the algorithms was judged based on the calculated speedup. From preliminary tests using a large dataset of two million values, there was a speedup of 2.6 when using a max of 6 threads in the thread pool.

College

College of Science & Engineering

Department

Computer Science

Campus

Winona

First Advisor/Mentor

Sudharsan Iyengar

Second Advisor/Mentor

Mingrui Zhan

Third Advisor/Mentor

Trung Nguyen

Start Date

4-24-2025 9:00 AM

End Date

4-24-2025 10:00 AM

Presentation Type

Poster Session

Format of Presentation or Performance

In-Person

Session

1a=9am-10am

Poster Number

45

Comments

No poster file

Share

COinS
 
Apr 24th, 9:00 AM Apr 24th, 10:00 AM

Comparing Performance of Parallel Implementation of Sorting Algorithms Versus Standard Implementations

In this research, sorting algorithms that utilize parallel programming were implemented to determine if parallel programming can be utilized to create an algorithm with a faster execution time. An implementation of the quick sort was developed that utilizes parallel programming via the Fork Join class in Java, and it was compared to the sequential quick sort algorithm. The comparison was made by using arrays of different sizes with unsorted integers. The arrays were sorted through both algorithms and parallel algorithms with various threads. The best algorithm and the best thread count were determined. The efficiency of the algorithms was judged based on the calculated speedup. From preliminary tests using a large dataset of two million values, there was a speedup of 2.6 when using a max of 6 threads in the thread pool.