NVIDIA and its partners
are rolling out a new option for engineering organizations or research shops
looking for a streamlined way to deploy high-performance computing (HPC) type
horsepower to run sophisticated, data-intensive applications.
Preconfigured Cluster harnesses NVIDIA's Tesla Graphics Processor Unit
(GPU) architecture in a ready-to-power-up, off-the-shelf solution designed to
allow research departments, engineering groups and IT shops to easily add GPU
computing capabilities to their existing data center systems. The cluster
system provides engineers with up to 30 times the performance of CPU-only
solutions in a smaller, more-efficient platform, delivering the necessary gains
to run such applications as computational fluid dynamics, molecular dynamics,
seismic processing and finite element analysis while at the same time lowering
overall energy costs, according to an NVIDIA spokeswoman.
Time on supercomputers is in high demand yet very difficult
to get, according to Andy Keane, general manager of the Tesla business unit at
NVIDIA. Engineers, scientists and researchers are increasingly in the market
for computational horsepower to run more sophisticated applications and
simulations, yet they frequently have to line up to get access to resources and
then wait while some sophisticated computations take weeks or even months to
The Preconfigured Clusters can easily create that kind of
HPC platform in existing data centers. Comprised of x86 CPU servers coupled
with the Tesla S1070 1U GPU systems, configurations start at 16 Teraflops of
performance, delivered by four Tesla S1070s, each containing four Tesla
10-series GPUs. Each system includes host servers, infiniband switches and cabling
and can be customized to suit individual customer needs.
Tesla GPU partners including Cray,
Colfax International, AMAX and Penguin Computing among others are offering Tesla Preconfigured Clusters.