Appro today announced that it has won the contract to deploy an interface data analysis and visualization cluster called “Graph” to Lawrence Livermore National Laboratory. The lab’s visualization and data analysis resources play a key role in the high performance computing ecosystem at LLNL. In this case, Graph will be used to visualization and data analysis duties in association with the extreme scale Sequoia machine.
To meet our current and future scientific computing needs requires a visualization cluster with enough memory to generate visualizations of simulation runs on our largest compute platforms as well as sufficient I/O rates for interactive analysis,” said Becky Springmeyer, Visualization group leader in LLNL’s Advanced Simulation and Computing Program.
Appro deployed four Scalable Units [SUs], totalling 110 TFlops of compute capability for Graph. The machine consists of 564 compute nodes, 2,256 processors/13,536 cores with 73TB of memory. Each compute node is based upon AMD’s Six-Core Opteron Istanbul processors clocked at 2.0Ghz. The fabric consists of 48 Flextronics 20Gbps DDR edge switches and two Voltaire DDR spine switches.
For more info, read their full release here.