Multi-GPU Computing for Achieving Speedup in Real-time Aggregate Risk Analysis

Stochastic simulation techniques employed for portfolio risk analysis, often referred to as Aggregate Risk Analysis, can benefit from exploiting state-of-the-art highperformance computing platforms. In this paper, we propose parallel methods to speedup aggregate risk analysis for supporting real-time pricing. To achieve this an algorithm for analysing aggregate risk is proposed and implemented in C and OpenMP for multi-core CPUs and in C and CUDA for many-core GPUs. An evaluation of the performance of the algorithm indicates that GPUs offer a feasible alternative solution over traditional high-performance computing systems. An aggregate simulation on a multi-GPU of 1 million trials with 1000 catastrophic events per trial on a typical exposure set and contract structure is performed in less than 5 seconds. The key result is that the multi-GPU implementation of the algorithm presented in this paper is approximately 77x times faster than the traditional counterpart and can be used in real-time pricing scenarios. Keywords-GPU computing; high-performance computing; aggregate risk analysis; catastrophe event risk; real-time pricing