GPGPU: Massively Parallel ComputingCUDA GPGPU COMPUTING GE Intelligent Platforms is implementing a family of rugged products based around the NVIDIA® CUDA™ architecture. With a vast resource of reference material and wide range of proven application use, CUDA is set to revolutionize the computing capability within the Mil/Aero application space. Across applications as diverse as SAR simulation, pattern recognition, object detection and cryptography, NVIDIA's customer base reports performance increases of between 10x and 500x! Using a GPU to augment the capability of a traditional CPU can therefore bring real size, weight & power (SWaP) advantages to high-end processing systems.
In addition to CUDA™ for massively parallel computing, the family will support PhysX® technology for exceptional realism in high-end graphics solutions, such as embedded training. With both 3U and 6U form factor families, GE Intelligent Platforms is well positioned to provide the processing power needed to satisfy next generation capabilities in radar, sonar, UAV and ground vehicle platforms. 3U CUDA FAMILYThe 3U CUDA family is based around the GRA111 OpenVPX-compatible card. With a wide range of single board computers and supporting ultility products, this family offers great performance and high flexibility in a small package. The GRA111 is based on the NVIDIA GT 240 GPU. It features 96 CUDA cores, 128-bit memory interface, and 16-lane PCIe Gen 2 interface.
A CUDA 3U Starter Kit is available for quick and easy start up of a CUDA-based system, either for evaluation and benchmarking or to develop a specific application. Get the CUDA 3U Starter Kit Datasheet Now! Using this family, it is possible to specify a multi-board system tailored to specific application requirements, or use the CUDA-enabled MAGIC1 off-the-shelf packaged solution.
The event will provide a unique opportunity to learn, collaborate and network. As a GTC 2010 sponsor, GE Intelligent Platforms is pleased to invite our partners and customers and offer a discount on registration through use of the code GXMGEP15. EXCLUSIVE OFFERGPGPU: Massively Parallel Computing Register here to be contacted by GE Intelligent Platforms Intelligent platforms sales for advance information on Increasing Functional Capability While Reducing Size, Weight & Power (SWaP) through the use of CUDA-enabled rugged products.
6U CUDA FAMILYThe 6U CUDA family is based around our AXIS Multicomputing hardware and software. With a range of 6U VPX cards it is possible to architect solutions comprising clusters of Intel and NVIDIA processors interconnected with high bandwidth fabric solutions. Data flow through the system can be optimized to produce systems which deliver huge volumes of data to the CUDA-capable NVIDIA GPUs to crunch. This family is ideally suited to medium- to high-end processing requirements within the Mil/Aero vertical market.
RESOURCES GE GPGPU information
NVIDIA CUDA documentation OpenCL documentation VPX information
Press releases
|
