Share |

The energy barrier

Pictured here is the Global Cloud-Resolving Model, represented on a 10,242-cell spherical geodesic grid. The model was calculated using the Green Flash. Image courtesy of NERSC.

It's widely acknowledged that high-performance computing systems are in trouble if they don't come up with a solution to their growing power needs. This is often used as an argument for why we all need to move to GPUs, which are more power efficient than traditional CPUs.

Here at iSGTW, we don't have a crystal ball. But we do have a spotlight we can shine on possible solutions. Although we haven't had a chance to address field-programmable gate array chips, or FPGAs, in this publication, we've been aware of some innovative work on the concept.

FPGAs are thought to have potential in this regard because, as the embedded chips that are used in mobile technologies such as cellular phones, they have always had to be designed with energy efficiency in mind. After all, no-one likes a power hungry cellphone that is constantly running out of battery power.

Apparently, we're not the only ones who have noticed FPGAs. KQED, a publicly funded radio station in Northern California, recently ran a short radio spot on the energy barrier and FPGAs. It's aimed at the general public, but it's a fun listen. You can read the article version or listen to it by clicking here.

Want to learn more about FPGAs and HPC? Check out these links:

Your rating: None Average: 3 (1 vote)


Post new comment

By submitting this form, you accept the Mollom privacy policy.