Appro will soon be rolling its latest cluster node design off the production line. The HyperPower series cluster combines a core platform of Intel Nehalem silicon with NVIDIA Tesla GPUs. According to VP John Lee, Appro has been looking at using GPUs for computational workloads for several years. It was only recenltly, however, that it made sense to incorporate their use into the Appro product line.
There’s always been a need in HPC to find creative ways to run code faster, and there’s always been an interest in specialized CPUs,” John Lee, vice president of advanced technology solutions for Appro, said in an interview.
The HyperPower cluster will support up to 304 scalar cores and up to 18,240 GPU cores in a single 42U/19″ rack. This equates to roughly 78TF of single precision compute. It will be marketed towards users in the federal government, academia and industry. For more info, read the full release here or the article on eWeek here.