Feature - Dash heralds new form of supercomputing

Dash, pictured here, is an element of the Triton Resource, an integrated data-intensive resource primarily designed to support UC San Diego and UC researchers. Image courtesy of San Diego Supercomputer Center, UC San Diego.

The first of a new breed of supercomputers was born this fall when computer experts combined flash memory with supercomputer architecture to create Dash.

Normally supercomputers are measured by how many floating point operations, also known as “flops,” they can complete per second. And at a peak speed of 5.2 teraflops, Dash wouldn’t even make the top 500 list, where the slowest speed is about 17 teraflops.

“But if you look at other metrics, such as the ability to do input/output operations, it would potentially be one of the fastest machines,” said Allan Snavely, the project leader for Dash. “Dash is going after what we call data-intensive computing, which is quite different from traditional high-performance computing.”

Traditional supercomputers are suited to calculating partial differential equations representing complex systems such as the weather, said Snavely. Dash’s flash memory enables it to read and write data 10 times faster than a typical supercomputer, which makes Dash perfect for processing huge data sets such as the human genome or large astronomical sky surveys.

Eventually researchers will be able to access Dash via TeraGrid. Until then, only researchers at Dash’s home institutions – the San Diego Supercomputer Center and the University of California-San Diego – plus a handful of external researchers can use it.

Meanwhile SDSC researchers are running tests on Dash to help determine what the next generation system should look like, said Snavely. “We’ll probably be building [the next system] over the next year.”

Dash, which was conceived as a prototype, is a hybrid system with four terabytes of flash memory and an additional three terabytes of DRAM. The new system will likely be larger and incorporate a larger proportion of flash memory.

As for the lack of recognition computers like Dash get, Snavely believes that’s about to change. “You’ll still see big flop machines like the Jaguar system, and those will still be used for the big mathematical calculations,” said Snavely. “But I think you’ll also start seeing data-intensive supercomputers. Maybe we’ll even see a top 500 list for supercomputers based on their input/output speed.”

—Miriam Boon, iSGTW