oa Massive Parallel Simulation of Motion of Nano-Particles at the Near-Wall Region in a Micro-Fluidics System

One of the major challenges in Computational Fluid Dynamics (CFD) is limitation in the available computational speeds, especially when it comes to N-body problems. Application of the graphics processing units (GPU) are considered as an alternative to the traditional CPU usage for some CFD applications to fully utilize the computational power and parallelism of modern graphics hardware. In the present work a Matlab simulation tool has been developed to study the flow at the wall-fluid interface at nano scale in a micro-fluidic system. Micro-fluidics has become progressively important in recent years in order to address the demand for increased efficiency in a wide range application in advanced systems such as “lab-on-a-chip”. Lab-on-a-chip devices are miniature micro fluidics (labs) to perform biological test such as proteomics, or chemical analysis/synthesis of very exothermic reactions. For simulation in such cases, it takes few hours to simulate a particular set of parameters even when run in parallel mode at the TAMUQ supercomputer. In order to accelerate the simulation, the Matlab program has been ported to C/C++ program that exploits the GPU's (Graphics Processing Unit) massive parallel processing capabilities. Results from both methods will be shown and conclusions be drawn.