GPU-based parallel generation method for stochastic models of two-phase media, electronic devices

Abstract

The invention discloses a parallel generation method of a random two-phase medium model based on a GPU and electric equipment. The method comprises the steps that a CPU reads initial distribution information corresponding to two-phase media; the CPU defines a target two-point probability function and a target linear path function; the GPU generates an initial random model; the CPU generates a current two-point probability function and a current linear path function, and a combined error is calculated; the CPU judges whether the combined error is smaller than a target value or not; if the combined value is smaller than the target value, the CPU draws the current initial random model, and the current initial random model is stored as the final two-phase medium model; if the combined error isnot smaller than the target value, the iteration step is executed repeatedly until the combined error is smaller than the target value, then the CPU draws the current initial random model, and the current initial random model is stored as the final two-phase medium model. The parallel generation method is high in computing speed, low in computing cost, low in hardware cost and convenient and easyto use.

Description

technical field [0001] The invention relates to the technical field of computer aided computing, in particular to a GPU-based parallel generation method of a random model of a two-phase medium and electronic equipment. Background technique [0002] Two-phase media, such as certain suspensions, porous media, or composite materials, are composed of two different types of discrete particles. In materials science we refer to a type as a phase. Two-phase media materials are obtained by randomly infiltrating particles of one phase and mixing them in another phase. Usually, we refer to the phase of the infiltrating particles as the second phase, and the infiltrating target as the matrix phase. In practical applications, the matrix phase is usually a fluid, a solid, or a space. The properties of such two-phase dielectric materials, such as electrical conductivity, elastic modulus or fluid permeability, depend on the distribution between the two phases and are usually also characte...

AI Technical Summary

Problems solved by technology

This type of parallel computing method has the following disadvantages: on the one hand, the computing efficiency of distributed computers is related to the communication efficiency between nodes or multi-threads, and excessive communication volume will reduce the parallel efficiency; on the other hand, the Hardware is expensive, and computing platforms are difficult to build and popularize

However, in the field of computational materials science, the random generation of random models of two-phase media still cannot use GPU-based parallel computing technology to improve computational efficiency.

Images