GE Intelligent Platforms has announced at the NVIDIA GTC (GPU Technology Conference) the 6U OpenVPX CUDA Starter Kit. Allowing systems developers to easily and cost-effectively acquire a complete, integrated, pre-tested hardware and software development environment, it is designed as a response to the rapidly growing interest in using GPGPU -- general purpose computing on a graphics processing unit -- technology for rugged, size/weight-constrained, deployed mission computing applications such as radar, sonar, image processing, signals intelligence and ISR (intelligence, surveillance, reconnaissance). By shortening the time-to-solution, the 6U CUDA Starter Kit can deliver customers a sustainable competitive advantage.

Concurrently, GE Intelligent Platforms is also announcing AXISLib-GPU, an advanced software environment that supports rapid development and debugging of CUDA-based applications.

"When compared to FPGA-based application development, CUDA has been demonstrated to be both faster and less expensive -- and now, getting started on exploiting the potential of GPGPU technology just got a whole lot easier," said Rob McKeel, vice president, Military & Aerospace Embedded Computing at GE Intelligent Platforms. "A single part number gets the developer everything required to begin work, eliminating the possibility of configuration error, minimizing delay and maximizing productivity."

CUDA is a parallel computing architecture, developed by NVIDIA for its GPUs, that is accessible to software developers through industry standard programming languages and makes the latest NVIDIA GPUs accessible for massively parallel computation. CUDA is ideal for deployed mission computing applications: performance improvements for certain types of application, such as radar, have been demonstrated to be 10x or better when compared to CPU-only solutions.

The 6U CUDA Starter Kit bundle provides all the necessary hardware and software to enable development of these demanding multiprocessing applications, including an IPN250 single board computer with an Intel Core2 Duo processor and an NVIDIA 96-core GT240 GPU; an NPN240 dual NVIDIA GT240 GPU card to support multi-GPU configurations; a 4-slot chassis and ancillary hardware including disk drives, rear transition modules, power supply and cables; and a preinstalled CentOS Linux start-up image. 

The GE OpenVPX 6U CUDA Starter Kit supports multiple fabric topologies within a single chassis, such that the developer does not need to define the system back plane requirements in advance of application development. This means that engineers can experiment with the system design before committing to a particular fabric interconnect for the final system, reducing risk and time to market.

A 3U OpenVPX CUDA Starter Kit is also available.