By GPU usage in the financial industry continues to grow. Previously on this blog we told you about Aqumin, a company that takes complex financial information and makes it actionable by using GPUs for analytics as well as data visualization Now we wanted to share an update about Hanweck Associates’ product, Volera, which uses parallel computing to tackle the fast-moving world of options analytics.
Financial analysts are constantly flooded with “messages” or options trade and quote data. Literally millions of messages are transmitted every single second, an overwhelming amount of data. Analysts and traders are expected to analyze these messages and make fast and smart investment decisions. Using computers to analyze this data is the only way to go, but the serial method of computation used by traditional CPU-based computers is simply not capable of efficiently handling this amount of information. Only by tackling the data in parallel, can financial analysts get timely information. Volera does just that.
The product first launched in 2007 and back then Volera was able to price the entire US options market in real-time using just 12 GPUs as opposed to 60 1U servers equipped with CPUs. With the massively parallel CUDA architecture of NVIDIA’s latest GPUs, Volera performs options calculations faster than ever before. Even though message rates in the US options market have tripled since 2007, Volera today achieves the same performance with just 8 Fermi-class Tesla GPUs.
In a recent interview with the analyst firm TABB Group, Jerry Hanweck, founder and principal partner of Hanweck Associates had this to say about GPUs in finance:
“GPU technology is incredible – It’s one of those rare times when a technology will completely change an industry.”
Check out Jerry’s full interview below:
Tagged: GPGPU, Supercomputing, Tesla
