On the issue of fuzzy timing estimations of the algorithms running at GPU and CPU architectures

Oleg Agibalov,Nikolay Ventsov,D Rudoy,V Murgul

doi:10.1051/e3sconf/201913501082

Abstract

We consider the task of comparing fuzzy estimates of the execution parameters of genetic algorithms implemented at GPU (graphics processing unit’ GPU) and CPU (central processing unit) architectures. Fuzzy estimates are calculated based on the averaged dependencies of the genetic algorithms running time at GPU and CPU architectures from the number of individuals in the populations processed by the algorithm. The analysis of the averaged dependences of the genetic algorithms running time at GPU and CPU-architectures showed that it is possible to process 10’000 chromosomes at GPU-architecture or 5’000 chromosomes at CPUarchitecture by genetic algorithm in approximately 2’500 ms. The following is correct for the cases under consideration: “Genetic algorithms (GA) are performed in approximately 2, 500 ms (on average), ” and a sections of fuzzy sets, with a = 0.5, correspond to the intervals [2, 000.2399] for GA performed at the GPU-architecture, and [1, 400.1799] for GA performed at the CPU-architecture. Thereby, it can be said that in this case, the actual execution time of the algorithm at the GPU architecture deviates in a lesser extent from the average value than at the CPU.

Highlights

The stochastic behaviour of bio-inspired algorithms makes it difficult to determine both effective algorithmic structures, and the choice of hardware architecture with which the search procedure will be implemented
Modern hardware architectures differ by the number of cores and the number of operations performed by a single core per time unit but they differ by the algorithms for organizing the computing process at the physical level [2]
In study [7], it was shown that the dependences of the genetic algorithms (GA) operating time on GPU and CPU architectures on the number of GA individuals, constructed only on the basis of maximum and minimum values, can significantly differ from similar dependencies constructed on the basis of averaged data

Summary

Introduction

The stochastic behaviour of bio-inspired algorithms makes it difficult to determine both effective algorithmic structures, and the choice of hardware architecture with which the search procedure will be implemented. Typical representatives of bio-inspired approaches are genetic algorithms (GA), currently used to solve a wide range of optimization problems [1]. Modern hardware architectures differ by the number of cores and the number of operations performed by a single core per time unit but they differ by the algorithms for organizing the computing process at the physical level [2]. With a large number of processed variables, the actual object of research is the issue of evaluating the effectiveness of using various hardware architectures to solve optimization problems using bio-inspired methods

Scope of research

The research issue

Findings

Conclusion