oneAPI: Quant Finance Made Easy

In quantitative finance Monte Carlo method is a popular approach when using GPUs. It is implemented using CUDA or OpenCL. The acceleration device use for other areas of the quantitative finance is slowed down due to the high cost of implementation and low rate of speed up.  Thanks to oneAPI a complex pricing using sophisticated models and methods can be easily implemented. Here we demonstrate how with minimal efforts pricing of the callable CMS spread swap using 2-factor Hull-White model can be efficiently implemented. The Hull-White model is an interest rate derivatives pricing model. This model makes the assumption that very short-term rates are normally distributed and revert to the mean. The Hull-White model calculates the price of a derivative security as a function of the entire yield curve rather than a single rate.

D&I used SYCL to code the existing model in the new framework. The benefits include: a) less code, b) increased transparency, c) best performance thanks to the optimized libraries. Usually quantitative libraries have a 7-10 years life span. During this time one generic code is written to cover all previously known cases. The generic code tries to incorporate all specifics of most models and payoff, making it too complicated. SYCL allows to put the model specific part into the kernel code and leave the parallel execution to oneAPIconstructs. This speeds up the model development.