Abaco: rugged XMC delivers 2.3 TeraFLOPS of peak performance and support for CUDA and OpenCL

Abaco Systems announced a powerful new rugged XMC graphics and video capture board. The NVP2102 takes advantage of the high performance of the NVIDIA Pascal P2000 GPU with its 768 cores and 4 GBytes of GDDR5 memory to deliver 2.3 TeraFLOPS of peak performance with support for both CUDA and OpenCL. It is the first Abaco video/graphics XMC product to offer high performance input and output.

Developed for deployment in the harsh and cramped conditions typical of battlefield platforms, the NVP2102 provides an optimum choice for all forms of graphics/video processing input/output. Typical applications include situational awareness, signals intelligence, ISR and radar, while the Pascal GPU’s high degree of parallelism also makes it an ideal fit for GPGPU applications including machine learning and autonomy.

The NVP2102 is designed for graphics and video generation, input and output, supporting four 3G-SDI inputs, capable of 1080p60; two 3G-SDI outputs, also capable of 1080p60; two DisplayPort 1.4 ports capable of 4K resolution @ 60Hz; and VGA.  Support for H.265 (HEVC)/H.264 (MPEG-4 AVC) encode/decode is provided.

For electronic warfare applications in which minimal latency is an essential requirement, the NVP2012 provides a solution not only by virtue of its direct interface to the host computer (as opposed to card-to-card connectivity) but also provides direct video capture direct to GPU memory – again, minimizing latency.

The board meets the MIL-STD-810G standard: in its conduction cooled form, it is specified to operate at temperatures of -40 to +85°C and in its air-cooled form it is specified to operate at temperatures of -40 to +70°C.

The NVP2102 also benefits from support for Abaco’s AXIS ImageFlex, an image processing and visualization toolkit enabling rapid development of high performance image processing, visualization and autonomy applications aimed at size, weight and power sensitive platforms. AXIS ImageFlex is focused on high performance GPU processing and graphics with interoperability with other programming paradigms such as OpenGL, OpenCL, CUDA and OpenCV.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.