"The benefits of running our Algorithmics solution on Linux are tremendous. In addition to the improved performance and scalability, we have experienced reduced hardware, services and implementation costs," says Pierre Gaye, deputy head of capital markets in the risk division at SociÃ©tÃ© GÃ©nÃ©rale.
A long-term client of Algorithmics using Algo Credit on the Mark-to- Future framework, SociÃ©tÃ© GÃ©nÃ©rale uses Algo Credit's ability to forecast the mark-to-market values of deals in the portfolio at a given time horizon in the future. From these distributions, the bank can calculate a variety of risk indicators, such as potential future exposures for different confidence levels. SociÃ©tÃ© GÃ©nÃ©rale is now simulating approximately half of the financial instruments within its credit exposure book on a Linux-based architecture, significantly enhancing the performance of the bank's Algo Credit implementation. Running half of the financial instruments on a grid of HP Proliant DL360 G4s with dual Intel Xeon 3.6GHz processors, SociÃ©tÃ© GÃ©nÃ©rale has achieved a 33% performance increase on their entire portfolio.
The final performance gain will be even higher when SociÃ©tÃ© GÃ©nÃ©rale moves the remaining credit exposure simulations to Linux over the coming months.
The time savings will allow SociÃ©tÃ© GÃ©nÃ©rale to calculate more accurate credit exposures within the overnight batch window by allowing the bank to add more scenarios to the simulation.
Algo Suite on Linux makes it possible for simulations to be spread across multiple CPUs on multiple machines. After the implementation is complete, the solution scales easily and allows clients to add additional computing resources without further implementation or configuration work required, thereby reducing the total cost of ownership.
"SociÃ©tÃ© GÃ©nÃ©rale's Linux-ported Algo Credit solution is an important part of its best practice approach to enterprise risk management, delivering the risk information the bank needs to proactively advance its business," said Michael Zerbs, Algorithmics' President and Chief Operating Officer. "The bank's success is indicative of the enhancements in performance, scalability, reliability and ownership cost reduction in total hardware costs delivered by the platform."