High Performance Computing Research Paper

This survey was conducted to examine the current usage and future trends in how High Performance Computing (HPC) is being used within financial services organisations. Firstly, the areas in which HPC is being used were examined. The only surprise here was that there were a couple of cases of it being used in the Treasury area within banks.

With regard to business processes, not surprisingly one of the biggest drivers for the growth of HPC in the future are the programmes that Banks have in place to address the challenge of the new ‘fundamental review of the trading book’ (FRTB) regulation.

We also identified areas such as in Monte Carlo vs Closed Form and Model and other optimisations where it is used to reduce calculation time and make the process more efficient. Here the Tier 1 banks are clearly ahead. In the area of Operating Systems used to run HPC we found a high usage of Windows, although Linux was still dominant. This was presumably due to the number of Quant libraries that are Windows based.

C++ is still the language of choice in the Tier 1’s (closely followed by Python) with C# being the most popular in the Tier 2’s. Hadoop is dominant in the Big Data Analytics and non-SQL space. There is very little sharing of HPC with other areas such as AI. In the Tier 1 banks the HPC capability is mostly ring fenced, whilst in the Tier 2 banks it is mostly shared.

Pour approfondir ce sujet, veuillez contacter notre spécialiste