Jay’s Weblog

We are now in the age of Petascale computing. On June 26, 2007, IBM unveiled Blue Gene/P, the second generation of the Blue Gene supercomputer that is capable of running continuosly at 1 petaflop. (1 petaflop = 1 Quadrillion (1 x 10¹⁵) floating point operations per second).

“The one-petaflop Blue Gene/P supercomputer configuration is a 294,912-processor, 72-rack system harnessed to a high-speed, optical network. The Blue Gene/P system can be scaled to an 884,736-processor, 216-rack cluster to achieve three-petaflop performance. A standard Blue Gene/P supercomputer configuration will house 4,096 processors per rack.”

Not to be left out, Sun Microsystem launched its own ultra-dense petascale system called Constellation. A petaflop capable installation is in the works at the Texas Advanced Computing Centre (TACC), University of Texas.

“The Ranger cluster will deliver 1.7 petabytes of storage using the Sun Fire X4500 data servers, the highest density available. Once completed, the TACC installation will consist of over 80 Sun Constellation System racks of computing power totaling over 15,000 quad-core microprocessors, all connected by Sun’s new high density, 3456-port InfiniBand switch. Sun Grid Engine will be used as a resource manager to dynamically allocate compute resources to applications.”

Some nice pictures of Constellation available on Jonathan Shwartz’s blog post and Josh Simmons’s blog post.

The 29th Top 500 List was released recently. This list of world’s fastest supercomputers was dominated by US research lab with IBM’s bluegene at LLNL taking the top spot for the forth time in a row.

The fastest supercomputer in Europe is the Barcelona Supercomputing Center (ranked 9) in Spain. This supercomputing center is home to the GRID superscalar project (GRIDs), which allows grid-unaware applications to be parallelised at runtime and at task level. This project has now been Open sourced, more information can be found on the GRID superscalar web site.

The majority of supercomputers on the list cater to scientific computing. However, an interesting observation is the London IBM Deep Computing Capacity on Demand (DCCoD) supercomputer (ranked 100), which is currently being offered on a pay-as-you-go basis to financial companies.

“The offering is targeted primarily at financial markets customers, who require additional, high-performance computing power to run intraday and post-trading analytics, for example. The IBM/Intel platform offers a fast, highly secure addition to companies’ computing infrastructure which can be used on a flexible basis. The solution can be purchased in increments as small as eight hours a day, five days a week.”

In my opinion, pay-as-you-go computing for financial services will be a hard-sell due to the numerous restrictions and regulations (e.g. SOX) that govern the finance and banking industry in terms of data protection and risk management.