Monday, April 4, 2016

Curtiss-Wright’s New Rugged Dual GPGPU OpenVPX™ Module is Industry’s First to Deliver 5 TFLOPS Performance for Deployed Radar, SIGINT, EO/IR Applications

Ashburn VA April 4, 2016 – Curtiss-Wright’s Defense Solutions division today announced its second generation rugged GPU module for deployed Radar, SIGINT and EO/IR high performance embedded computing (HPEC) applications. The new CHAMP-GP3 GPGPU (general purpose graphics processor unit) Application Accelerator is the industry’s first 6U OpenVPX module to bring the extensive floating point processing power of NVIDIA®’s latest generation of Maxwell architecture class GPGPUs to the embedded aerospace and defense market. The board features dual 1536-core Tesla M6 MXMs high performance computing devices to provide an unprecedented 5 TFLOPS of compute performance in a single slot. Designed for compute-intensive applications that require extreme processing, the CHAMP-GP3 is the first COTS module to provide a size, weight and power (SWaP) optimized solution for system designers who seek the unmatched super computing capability of a Tesla GPGPU to address applications that can’t be satisfied by aggregating multiple microprocessors. In conjunction, Curtiss-Wright also announced that it has been selected by NVIDIA to become a member of its NVIDIA Partner Network program. 
The CHAMP-GP3 is designed for use in ISR and EW applications that require TFLOPS of accelerated processing. Thanks to its support for the popular CUDA™ and OpenCL programming languages, system designers using the CHAMP-GP3 can develop new applications in a fraction of the time required by FPGA-based solutions.  The module is ideal for addressing the massive amounts of data generated by modern Radar, SIGINT and EO/IR sensors and provides unparalleled HPEC performance in cross-cueing applications.
“We are very pleased to introduce the embedded industry’s highest performance OpenVPX GPU processor, the new CHAMP-GP3, which results from our exciting new alliance with NVIDIA as a member of their NVIDIA Partner Network,” said Lynn Bamford, Senior Vice President and General Manager, Defense Solutions division. “In recent years, we have seen the use of GPUs in ISR and HPEC applications leap from the experimental stage to the implementation stage. These SWaP-sensitive programs are hungry for as much shear processing power as they can fit into a single board slot. With 5 TFLOPS of floating point performance the CHAMP-GP3’s dual Tesla GPGPUs deliver the performance ISR and HPEC customers have been seeking.”

No comments:

Post a Comment