FPGA for financial applications

FPGA unit


Field Programmable Gateway Arrays (FPGAs) are all the rage in some areas nowadays, and for a good reason. They allow creating low level high performance hardware components that allow very fast execution of certain tasks. When comparing to the ubiquitous regular CPUs, found in standard servers and desktops, they enable the development of custom logic at much lower level. Their key advantage is the ability to provide parallel execution of routines that can be done only sequentially on a single CPU. And when compared to GPUs, the so popular graphic cards

Applied to the world of financial algo trading / HFT software development, FPGAs can provide help with the following routine applications:

  • Parsing of FIX or other incoming messaging format
  • Parsing and filtering of huge volumes of market data
  • Running mission critical parts of the risk or pricing models

These and many other possible applications can give you the advantage you need in order to bit your rivals and achieve better executions in the ever more competitive markets.

Real world financial applications

A number of major investment banks, including JPMorgan, Deutsche Bank, Credit Suisee and others, have been working on solutions using FPGAs in recent years. They are usually quite secretive with regards to the details of their applications, so only glimpses of what is going on inside those institutions reaches us. However, they do make massive investment in those programs and those who wish to compete with them on the same field should pay attention.

Potential costs

FPGA hardware used to be very expensive. Thanks to advances in technology and mass production this is no longer the case. At the time of writing an entry level system can be acquired for as low as $5,000. The actual development costs will, of course, depend on the nature of the application being developed.

Our expertise

There is, of course, a price to pay for the usage of this technology. The development for FPGAs, done at much lower level than traditional software development, is usually slower, requires knowledge of embedded programming techniques and is not something you find in your average financial coder. Fortunately, members of our team have decades worth of embedded industry experience among them, including programming for FPGAs. We would be glad to discuss your particular needs further and to see how our knowledge and expertise can help your particular needs. Please do not hesitate to contact us.