THE EFFICIENCY EXPLORATION OF PARALLEL WAVE ROUTING ALGORITHM WITH GPU COMPUTING COMPARED TO CPU
Abstract
The present paper concerns the issues of speeding up the execution time of the modified reverse wave routing algorithm in a software-defined network of large size. The parallel version of the algorithm is executed on the predefined network sizes with the same edge probability on a multi-core CPU and GPU separately, partly on a multi-core GPU and partly on a multi-core CPU. The exploration results of the parallel algorithm help to define the most suitable way of algorithm computing in networks of different sizes.
Keywords: modified inverse wave algorithm, CPU, GPU, software-defined network.
Fig.: 4. Tabl.: 3. Bibl.: 3.