VPX6-4955

Providing up to 22.4 TFLOPS, this rugged VPX6-4955 GPGPU board features a chip-down design to meet the requirements of rugged military and aerospace environments. The board provides a top tier module for intense processing and artificial intelligence (AI) in high-Performance Embedded Computing (HPEC) systems. The VPX6-4955 delivers incredible processing power from two NVIDIA Quadro Turing TU104 GPUs, providing a remarkable 6144 CUDA cores for parallel processing, 768 Tensor cores for deep learning neural network training, parallel processing, AI inference, and a 50% improvement in performance per CUDA core compared to the previous Pascal generation.

In addition to 6144 CUDA cores for parallel processing, each TU104 also features 768 Tensor Cores for dedicated AI inference and 96 ray-tracing (RT) cores for superior rendering speeds. What’s more, the Turing architecture’s Tensor Cores accelerate the tensor/matrix computation used for deep learning neural network training and inference operations, answering the growing demand for artificial intelligence and high-performance processing in deployed EW and ISR applications.

This board includes eight DisplayPort 1.4 outputs, which support High Dynamic Range (HDR) video at resolutions of 4K at 120 Hz or 5K at 60 HZ with 10-bit colour depth. The GPUs on the VPX6-4955 also features an improved NVENC/NVDEC accelerator for HEVC (H.265) and AVC (H.264) with up to 8K encode resolution and B-frame support.

Bring new levels of performance to your ISR and EW applications with the VPX6-4955:

The NVIDIA Turing architecture provides a 50% improvement in performance per CUDA core compared to the previous Pascal generation. Its redesigned memory path combines shared memory, texture caching, and memory load caching into one unit, resulting in twice the bandwidth and over twice the capacity for L1 cache and common workloads than the previous Pascal generation.

Moving data quickly and efficiently is paramount for high-speed GPUs. The VPX6-4955 uses GDDR6 memory, which provides twice the bandwidth of the GDDR5 memory used in previous generations. Incorporating a PCIe switch, this module is configurable for compatibility with various OpenVPX slot profiles.”

Hardware Accelerated Video Encode/Decode

Featuring the latest generation encode/decode hardware acceleration engine, the VPX6-4955 adds support for HEVC (H.265) 8K encoding at 30 fps and B-Frame support. This new engine achieves up to a 25% bitrate saving for HEVC and up to 15% bitrate saving for AVC (H.264) while producing real-time 8k and 4k encoding without burdening the CUDA cores.

Like previous versions of the encoding engines, NVENC supports CBR and VBR rate control, programmable intra-refresh for error resiliency. The Turing GPUs introduces new hardware functionality for high-performance computing of the relative pixel motion (optical flow) between images. These algorithms effectively handle frame-to-frame intensity variations and track the true object motions much more accurately than the traditional Motion-Estimate mode of NVENC.

VPX6-4955 is ideal for high-performance radar, SIGINT, EO/ IR, sensor fusion, processing and display, autonomous vehicles and SWaP-constrained ISR and EW applications requiring very high GPGPU processing power.