Description:

Time is money, the saying goes. And nowhere is this more true than in the world of finance. GPU supercomputing, with its ability to speed up many types of financial calculations by orders of magnitude (x10, x100, ...) looks set to do very well in finance. Indeed, it has already made significant inroads into financial analytics and risk in what looks like a silent revolution on Wall Street. The revolution has gone unnoticed for a while, but more and more financial players are seeing--and reaping--the benefits of GPU supercomputing.

GPU supercomputing in finance is about doing things faster than ever before. Trading is already mostly real-time, because it pays to be there first, ahead of the other guy. And even those aspects of finance that are not yet real-time (currently risk and compliance often lag behind trading), they are moving to real-time as well. In fact, the whole world seems to be moving to real-time; finance will simply get there first because the incentives are so great.

But with all this newfound speed there are dangers. Is faster really better for financial markets as a whole? Does faster mean more abundant liquidity, or less? More volatility, or less? Do the regulators need to go real-time in what amounts to a regulatory compliance arms race? And should they equip themselves with the same tools? GPU supercomputing looks to be sufficiently powerful and cheap--and orders of magnitude speedups must inevitably change the rules of the game--so this presentation will touch upon some wider policy-related issues.

This talk will give an overview of GPU supercomputing in finance: how it is being applied now, and where it's going; what sorts of problems in finance that GPUs are a good match for, and what problems are not, and how it might profitably be applied in your business.

About Andrew Sheppard

Andrew is a financial consultant with extensive experience in quantitative financial analysis, trading-desk software development, and technical management.

Most recently, from 2006 to 2009, Andrew was the Chief Technology Officer and Chief Quantitative Analyst at a New York multi-strategy hedge fund. He also was the manager of an innovative software company based in London that was owned by the hedge fund but run independently. For the last two years, Andrew has been an active developer of GPU (CUDA) massively parallel software in C/C++ for realtime financial trading and risk. Andrew entered finance after conducting scientific research at Oxford University, Caltech's Jet Propulsion Lab, (Pasadena, California) and the Berkeley Space Sciences Lab (Berkeley, California), where he worked on earth and planetary remote sensing probes.