JP Morgan risks all on a supercomputer

While the rest of the finance industry has been reeling from problems based on the automation of its decision making processes, Financial house JP Morgan is allowing a supercomputer to start making some of its important decisions.

The outfit is running its run risk analysis and price its global credit portfolio using a High Performance Computing (HPC) system developed by Maxeler Technologies.

The system is based on a Field-Programmable Gate Array (FPGA) technology that would allow it to run complex banking algorithms faster.

The financial house uses mainly C++ for its pure analytical models and Python programming for the making things happen. Under the new Maxeler system, the C++ code has been replaced mostly by Java.

The outfit has been trying to work out why it failed to spot suspicious Madoff trades, when it was possible to see his plans by just Googling the bloke.

Part of the problem, JP Morgan claims, is that it would take eight hours to do a complete risk run, and an hour to run a present value, on its entire book. If anything went wrong with the analysis, there was no time to re-run it.

Now the supercomputer can do it all in about 238 seconds.

Stephen Weston, global head of the Applied Analytics group in the investment banking division of JP Morgan, said that if the company can compress space, time and energy required to do these calculations then it has business value.

Rather than looking at a company’s risk based on the previous day, it can do the whole lot, more or less in real time.

This speed can identify potential problems and try to deal with them in advance. Now it is able to spot its exposure to problems such as the Irish or Greek bank problems, which was unthinkable before.

Instead of using existing standard multi-core machines, JP Morgan used FPGA technology which can ‘pipeline’ instructions. This means calculations can be executed very quickly by breaking down calculations into simple components that can then be built into ‘pipelines’.

By being super pipelined it can do a huge volume of calculations more than it can with a ‘traditional’ CPU.

The project took JP Morgan around three years, and the bank is now looking to push it into other areas of the business.