Abstract
Modern Graphic Processing Units (GPU) provide in combination with a very fast Video Random Access Memory (VRAM) very high computational procedure, outrunning the conventional combination of a Central Processing Unit (CPU) and Random Access Memory (RAM) in terms of parallel computing and calculation. Within this work a concept for parallel application of the CPU/GPU is presented which combines the approach for processing and managing of large amounts of data. The computer algebra system (CAS) Wolfram Mathematica is used for numerical calculation of a large Finite Difference Model (FDM). The CUDA-link feature of Mathematica was used to achieve a parallel working environment with a parallelized computation on available CPUs with a parallelization of calculations of Nvidia GPUs at the same time. An advanced desktop computer system was setup to use a high-end desktop CPU in combination with four TITAN GK110 Kepler GPUs from Nvidia. It will be shown, that the calculation time can be reduced by using sharedmemory and an optimization of the used block and/or register size to minimize data communication between GPU and VRAM. Results for diffusion, stress field and deformation field for a deformation sample will be shown, which is numerically calculated from crystal plasticity, with over four million of FDM elements being calculated by each of the four used graphic cards. It will be clearly shown, that the overall calculation time is strongly depending on the storage time for the amount of data, both for the final result and as for the intermediate results for the different numerical increments. Nevertheless, a promising application of parallel computing for research in the field of materials science is presented and investigated, showing the possibilities for new approaches and/or more detailed calculations in a reasonable time.