A review of purpose-built accelerators for financial services

This article provides an overview of purpose-built accelerators (PBAs) like GPUs and their applications in the financial services industry (FSI). It covers the background and history of accelerated computing, the technology behind PBAs, the software ecosystem for accelerators, and their applications in FSI use cases like data parsing, analytics, machine learning, and low-latency applications.

Specifically, the article covers:

• Background on accelerated computing and the rise of PBAs
• Review of the core PBA technologies like GPUs, FPGAs, and ASICs
• AWS accelerator hardware offerings like P5, Inf2, and Trn1 instance types
• Accelerator software like CUDA, Neuron SDK, and open-source efforts
• Applications of PBAs in FSI use cases like data parsing, Monte Carlo simulations, risk calculations, generative AI, time series prediction, and low-latency trading
• Key takeaways and considerations for business leaders regarding PBA adoption and strategy