How Combining AI with High Performance Computing (HPC) Could Transform the Finance Industry

From risk compliance to fraud detection, the banking sector is set to benefit from the AI-HPC combination.

Key Takeaways

  • Increasing data demands mean that more sectors, including the FSI, are moving to high performance computing (HPC).

  • Merging Artificial Intelligence (AI) with an HPC infrastructure could bring massive benefits for the banking industry.

  • The key to combining AI and HPC is scalability.



The increasing amount of data being generated means that ageing IT infrastructure can no longer cope with modern workloads. As a result, businesses are increasingly turning to high performance computing (HPC). This involves aggregating computing power across multiple servers, resulting in higher performance than you would get from a traditional PC.

HPC was previously only used in areas where high levels of computation were required, including engineering, academic institutions, governments, and the military. However, the ever-increasing amount of data, combined with disruptive technologies like Artificial Intelligence (AI) means that HPC is becoming increasingly essential for complex workloads and data-intensive computing in enterprise. Cloud computing and multi-core processors have made HPC more accessible and as a result, the use of HPC is gradually spreading to new industry sectors. This is helping businesses in the digital era to increase processing speeds and meet modern data demands. One sector that could see huge benefits from HPC is the Financial Services Industry (FSI).

“The emerging AI community on HPC infrastructure is critical to achieving the vision of AI: Machines that don’t just crunch numbers, but help us make better and more informed complex decisions.”

Traditionally, HPC tends to focus on solving a small number of big problems, but in the case of FSI, workloads are more likely to include a vast number of small calculations. For example, Monte Carlo simulations can be used to enable banks to see all the possible outcomes of their decisions and assess risk accordingly.

Going beyond the capabilities of a regular PC, HPC enables financial organisations to get access to information faster, run applications more efficiently, analyse data more quickly, and streamline processes. Along with Monte Carlo simulations helping with risk assessment, HPC could also aid banks in fraud protection, as well as compliance with ever-changing banking regulations.

Some banks are already running Proof-of-Concept experiments (PoC) using HPC, and this is something we are likely to see more of in the coming months. Because it improves computing capabilities, HPC also enables businesses to make use of new innovations such as Internet of Things (IoT) technology and Artificial Intelligence (AI). Incorporating AI into an HPC environment gives organisations the capability to scale to accommodate emerging workloads. Scalability is the key to the integration of AI and HPC, and the two technologies are set to become increasingly intertwined.

“The emerging AI community on HPC infrastructure is critical to achieving the vision of AI: Machines that don’t just crunch numbers, but help us make better and more informed complex decisions,” said Pradeep Dubey, Intel Fellow, and Director of Parallel Computing Lab in a report.

Intel® Xeon® processors are optimised for HPC while the Intel® Scalable System Framework (Intel® SSF) is designed to help businesses address the workload requirements of AI in an HPC environment, at a lower cost and with less effort. This scalable HPC framework is designed to eliminate data bottlenecks and helps organisations to make use of new technologies like AI with little effort or risk.

Intel® Omni-Path Architecture (Intel® OPA) is one of the building blocks of Intel® SSF and delivers high bandwidth, high message rates, low latency, and high reliability. This enables fast communication across multiple nodes or servers in an HPC system, reducing the time taken for processing data. Intel® OPA enables business to scale cost-effectively, from small HPC clusters to larger clusters containing 10,000 nodes or more.

As the data revolution continues, FSI businesses must adapt to cope with more complex data. Moving to updated HPC infrastructure will enable them to tame the data deluge and accommodate advanced AI workloads.

For More Information: