Efficient bulk-parallel algorithms for data recovery in multidimensional case

A data recovery problem in multidimensional case by given values in the uniform nodes is considered. An efficient computational data recovery algorithm for a bulk parallelism model is proposed. The algorithm is based on transformations applied to summatory on V. A. Steklov’s kernels operator. The dependency between computational complexity and the problem parameters is shown. A software implementation of the algorithm is carried out in Open CL standard widely used for general-purpose computing on graphic processor units. Examples of surfaces reconstruction by the software are given.

Authors: A. R. Liss, E. I. Sergeeva, G. Yu. Puerov

Direction: Informatics, Computer Technologies And Control

Keywords: data recovery, parallel computing, bulk-parallelism, Open CL, graphic processor unit (GPU)

---

View full article