Differential evolution (DE) is a powerful population-based stochastic optimization algorithm, which has demonstrated high efficacy in various scientific and engineering applications. Among numerous variants of D
(1), (2), we first implemented Algorithm 3, Algorithm 1 using the CUDA Toolkit v2.1. We then translated the CUDA code into C++ code by replacing all kernel invocations with loops and removing unnecessary elements (such as references to shared memory, which does not exist on a CPU). We ...
Mahafzah BA (2013) Performance assessment of multithreaded quicksort algorithm on simultaneous multithreaded architecture. J Supercomput 66(1):339 Article Google Scholar Szgyi Z, Trk M, Pataki N (2011) Multicore C++ standard template library in a generative way. In: Proceedings of the Third Wor...
In the present approach, the evolution of quasi-particle objects is assigned to different GPU threads. In the case of multiple GPUs, the algorithm assigns to each GPU a subset of energies to be simulated; within this approach, it is not needed to share memory across devices and each CPU ...
analytical programmingspectra analysisCUDAevolutionary algorithmdifferential evolutionparallel implementationsymbolic regressionIn this paper we discuss a method useful for spectra analysis 鈥 analytical programming and its implementation. Our goal is to create mathematical formulas of emission lines from spectra,...
The model is solved numerically using a finite difference method while the time evolution is discretized by using explicit Euler scheme. For reduction, the elapsed time, the parallel algorithm based on Graphical Processing Unit (GPU) with CUDA was used. This study compares the elapsed time of ...
Combining Lattice Boltzmann method and genetic algorithm to optimize the layout of artificial floating islands in river network in China 2023, Environmental Science and Pollution Research A GPU accelerated study of aqueous humor dynamics in human eyes using the lattice Boltzmann method 2023, Mathematical...
This work proposes to study a parallel multi-objective algorithm, the multi-objective version of Differential Evolution (DE). The generation of trial individuals can be done in parallel, greatly reducing the overall processing time of the algorithm. A novel approach to parallelize this algorithm is...
3.2. Shared memory caching algorithm While modern GPUs provide peak computing performance of several trillion floating point operations per second (TFLOPS) the efficiency of real world applications is often much lower. For many applications the memory bandwidth and latency are a bottleneck, which can ...