Michael Feldman wrote an interesting article this week over HPCwire on the new ATLAS cluster at the Max Planck Institute for Gravitational Physics in Germany. Michael’s piece does a good job of summarizing the interesting work this machine will do, so check out his piece for that (I’ve always said that if I’d have been smart enough I would have been a gravitational physicist instead of an engineer).
What struck me about Michael’s piece is that the machine — a 32 TFLOPS Xeon-based cluster — is interconnected with Woven’s fancy GigE switch infrastructure. It turns out that the gravitational wave hunt this machine will go on is very parallelizable, but I predict that we’ll start to see more of Woven’s equipment used in situations that are less specialized
Another attraction of the Woven technology is its ability to dynamically determine the optimal path for the data. The vSCALE chip in the switch is constantly monitoring latency of the active and alternative paths in the Ethernet fabric. If it finds an alternative path with lower latency, the hardware redirects traffic to take advantage of the faster route. This is especially advantageous when all the nodes are accessing both central storage and local disks on the other nodes.
Check out the related stories in the links on the article page for some of the back story on Woven.