A Detailed Account of An Efficient Acceleration of Solving Heat and Mass Transfer Equations in Capillary Porous Radially Composite Cylinder with Different Kinds of Boundary Conditions Using Programmable Graphics Hardware

A Detailed Account of An Efficient Acceleration of Solving Heat and Mass Transfer Equations in Capillary Porous Radially Composite Cylinder with Different Kinds of Boundary Conditions Using Programmable Graphics Hardware

High attempts have been made to replicate a variety of scientific phenomena in engineering fields using the latest developments in computational technology. One such case is the simulation of heat and mass transfer in capillary porous media, which, in the study of a number of applications in science and engineering applications, is becoming increasingly important. This method of numerical solution of heat and mass transfer equations is, however, very time-consuming for capillary porous media. This paper therefore uses one of the acceleration methods developed in the graphics community that takes advantage of a graphical processing unit ( GPU) that is applied to the numerical options of such equations for heat and mass transfer. The programming model of the Navaid Compute Unified Computer Architecture (CUDA) provides a correct approach to applying parallel computing to the graphical processing unit framework. This paper indicates a true performance improvement when solving the equations of heat and mass transfer for capillary porous radically composite cylinders with many kinds of boundary conditions, running on GPU numerically. In a capillary porous cylinder, this heat and mass transfer simulation is carried out using the CUDA platform on the nVidia Quadro FX 4800 graphics card. We introduced numerical solutions using GPGPU ‘s highly parallel computing capabilities on nVidia CUDA. In the field of numerical solution to heat and mass transfer, we have demonstrated that GPU can operate significantly faster than CPU. Experimental findings for capillary porous radially composite cylinder show that our GPU-based implementation shows a substantial improvement in efficiency over CPU-based implementation and about 7 times the maximum speeds observed.

Author(s) Details

Dr. Hira Narang
Computer Science Department, Tuskegee University, Tuskegee, Alabama, USA.

Dr. Fan Wu
Computer Science Department, Tuskegee University, Tuskegee, Alabama, USA.

Abdul Rafae Mohammed
Computer Science Department, Tuskegee University, Tuskegee, Alabama, USA.

View Book :- https://bp.bookpi.org/index.php/bpi/catalog/book/307

Editor 251News

leave a comment

Create Account



Log In Your Account