BOPM implemented on a GPU-architecture
Department of Mathematics, Analysis and Applied Mathematics, Disciplinary Domain of Science and Technology, Mathematics and Computer Science, Uppsala University
Uppsala University, 2011
@article{furst2011bopm,
title={BOPM implemented on a GPU-architecture},
author={F{"u}rst, K.},
year={2011}
}
We used the Binomial Options Pricing Model (BOPM) implemented on a Graphics Processing Unit (GPU) to calculate the value of European and American options, of both put and call type. The advantage of using a GPU over a CPU is that a GPU has many more processing-cores than a CPU and can perform more calculations simultaneously. This, of course, also implies limitations. The result was compared with both the exact value and with the value obtained using a Central Processing Unit (CPU). The CPU-code was implemented using C++ and the GPU-code was implemented using NVIDIA’s programming language CUDA. It turns out that the value obtained using the GPU is for the most part almost consistent with both the exact value and the value obtained using a CPU.
December 28, 2011 by hgpu
