The Appro HyperBlade Cluster at Performance and Architecture Laboratory (“PAL”) at Los Alamos National Laboratory, reached a Linpack benchmark performance of 3.4 TFlop/s “teraflops” or trillions of floating point calculations per second. This Appro HyperBlade Cluster is placed in the one hundred and thirty-second position on the current TOP500 list. The Appro HyperBlade Cluster solution is powered by Dual-Core AMD Opteron(TM) processors configured in 2-socket 256 nodes for a total of 1024 cores. The cluster is running Linux and is connected by a Voltaire InfiniBand switch. This cluster reflects the growing trend in supercomputing toward standards-based systems configured in clusters to address sophisticated computing needs. For a cluster picture, go to http://www.taborcommunications.com/hpcwire/features/topsys/appro_pal.htm
Link: Yahoo PR News
