NVIDIA has announced the availability of preconfigured Tesla GPGPU clusters that are ready to go off the shelf to improve the performance of data centers for researchers and IT managers. The Tesla Preconfigured Cluster provides 30 times the performance of CPU-only solutions and allows IT managers to deploy smaller and denser servers.

The clusters consume significantly less power than traditional server clusters according to NVIDIA for a significantly savings in power and cooling bills. NVIDIA says that the Tesla Preconfigured Clusters include x86 CPUs coupled with Tesla S1070 1U GPU systems.

Configurations have at least 16 teraflops of performance with four Tesla S1070s each housing four Tesla 10-series GPUs. The systems all include host servers, infiniband switches, cabling and are customizable to meet user needs.

Via NVIDIA