International Monetary Fund adopts grid computing

High-performance computing expected to help analyse financial risks

The IMF's analysis and simulations structure could not cope with increasing data processing requirements. Pic credit: Alan Cordova

The International Monetary Fund (IMF) is implementing high-performance grid computing to cope with the increasing demand for risk analysis and financial modelling scenarios.

The improvements in the IMF platform are intended to boost performance and efficiency of complex financial applications in a cost-effective manner. Such models rely on simulations that were previously run on individual workstations and servers, which stalled productivity due to the resulting latency and computation run-time.

Increasing workloads and the limitations of the current set-up prompted the institution to adopt the on-demand high-performance computing system, which is hoped to provide a tenfold boost in processing time for economic modelling computations.

Under the new system, complex economic computations are distributed across under-utilised smaller computers within the IMF network, allowing calculations to complete faster and in a more predictable fashion.

“The IMF required a proven grid computing solution that would better utilise our existing compute resources to reduce the time required to run increasingly complex calculations,” said Jean Salvati, IT officer at the IMF.

Salvati said the new system will enable “significant response time improvement” for the IMF’s economists and will be “instrumental” as the body develops new modelling and simulation tools and in supporting the organisation’s role of monitoring financial institutions worldwide.

Platform Computing is providing the software.