SCIENCE

Appro (http://www.appro.com) has announced the availability of NVIDIA Tesla M2090 GPU computing capability for the Appro Tetra and GreenBlade server platforms and Appro Xtreme-X and HyperPower Supercomputing solutions.

Based on the CUDA architecture codenamed “Fermi”, the NVIDIA Tesla M2090 GPU Computing Module delivers approximately 10X higher HPC application performance than the latest multi-core CPUs. The Tesla M2090 GPU features many "must have" HPC capabilities including 512 computing cores per GPU at 665 gigaflops of double precision performance with ECC memory for uncompromised accuracy and reliability and an L1/L2 cache for improved bandwidth and data sharing.

The Appro Supercomputing solutions powered by the Tesla M2090 GPUs offer scalable and flexible building block architecture based on the Appro GreenBlade platform supporting up to 5 dual CPU server blades with 5 GPU expansion blades combinations in a 5U server form factor. It delivers up to 5,120 GPU cores per system. Also, Appro offers the Appro Tetra server platform with superior density based on a rack mount platform integrating 2CPU’s and 4GPU’s in one server while supporting up to 2,048 GPU cores per server. Appro Supercomputing solutions based on Tetra or GreenBlade platforms are pre-configured and pre-integrated including cluster and management tools with HPC professional services and support. These systems are outstanding compute density solutions for datacenter deployments and are ideal for seismic processing, biochemistry simulations, weather and climate modeling, signal processing, CAE, CFD, and data analysis.

“Appro is proud to collaborate with NVIDIA for this Tesla GPU upgrade”, said John Lee, VP of Advanced Technology Solutions of Appro. “Working together, Appro and NVIDIA provide higher application performance with energy efficient capabilities to meet the demands of many high-performance parallel computing users.”