high performance computing on graphics processing units: hgpu.org

Advanced applications of CFD for multiphysics modelling of electrokinetic, capillary, turbulent and rarefied hypersonic flows is discussed in this paper. Due the complexity of the geometry involved and the underlying physics associated with the phenomena to be studied, multiphysics study requires enormous computational resources. The CFD computations are performed within a parallel environment for accelerating solution time by distributing the processes on many nodes in a clustered supercomputing distributed parallel environment or by using the multi core CPU in a single node. Many CFD applications today can be accelerated significantly by using GPUs as the compute engine instead of the conventional CPUs. The current study investigates the significant improvements in the CFD simulations using multi core central processing unit (CPU) and graphical processing unit (GPU).