Scalability and performance are critical to successful risk management. Usually the more risk variables a system can model, the more accurate the predictive results. Yet more variables can lead to greater processing time. Therefore in the past many risk professionals have chosen to run simpler models with shorter processing times. With grid computing, however, solutions such as Ambit Risk & Performance can run complex models with more variables, giving accurate results in less processing time.
SunGard and Intel have been working on distributed computing with Ambit Risk & Performance in a grid environment since 2007. The 30% increase in computational performance came when the Ambit solution was implemented across a 96-core grid of Intel Xeon 5500 series servers. The percentage increase is taken in comparison to previous benchmark studies.
Paresh Pattani, director, software and services group, Intel, commented: âRunning risk simulations 30% faster without the need for code changes can lead to a competitive advantage. In these SunGard tests, the Intel microarchitecture, codenamed Nehalem, has demonstrated its ability to deliver improved performance and power efficiency for financial services applications.â
Andrew Chang, chief technology officer for SunGardâs Ambit business unit, commented: âToday, risk professionals want to perform ever more complex stochastic modeling, and they want to see the results even faster. It is therefore critical that solutions like Ambit Risk & Performance - which covers asset & liability management, market risk, credit risk, operational risk and economic capital â can harness grid computing to continually improve performance. We are very pleased with these latest results.â