Parallel Approaches Performance Evaluation Using Curves B-spline

• Vitor David UFRRJ
• Douglas Araujo UFRRJ
• Ubiratam de Paula UFRRJ
• Marcelo Zamith UFRRJ

Abstract

Given the importance of algorithm parallelization for performance, this paper proposes the performance analysis for different levels of parallel approaches allowing to conclude the best approach for b-spline curves. Performance was evaluated using different levels of parallelization, through instructions, with vectoring, and through cores, through threads. Our results demonstrated that a nucleus approach obtained very good performance while the vectorization did not gain.

Keywords: Parallel Programming, Interpolation, B-Splines

References

Bramas, B. (2017). A novel hybrid quicksort algorithm vectorized using avx-512 on intel skylake. International Journal of Advanced Computer Science and Applications (IJACSA), 8.

Buss, S. R. (2003). 3D computer graphics: a mathematical introduction with OpenGL. Cambridge University Press.

Holewinski, J.; Ramamurthi, R. (2012). Dynamic trace-based analysis of vectorization potential of applications. ACM SIGPLAN Notices, 47:371–382.

Stock, K.; Pouchet, P. S. (2012). Using machine learning to improve automatic vectorization. ACM Transactions on Architecture and Code Optimization (TACO) - Special Issue on High-Performance Embedded Architectures and Compilers, 8(50).

Wolberg, G. (1990). Digital image warping, volume 10662. IEEE computer society press Los Alamitos, CA.