Capabilities

Embedded Video Graphics, GPGPU & Computing Solutions

Using the latest GPUs available from NVIDIA, our COTS video graphics, capture, and GPGPU processing cards are designed to support edge processing in high-performance embedded (HPE) systems. The Condor line of 3U and 6U OpenVPX, XMC, and PCI Express form factor graphics cards are designed with high-speed memory, high bandwidth interconnects, and can be configured to meet various video formats.

Our GPGPU solutions are available in several GPU compute levels, supporting NVIDIA® Ampere, Turing™, or Pascal™ GPU based architectures. NVIDIA GPUs are packed with advanced capabilities including CUDA™ Cores, Ray Tracing Core, Tensor Cores, and CUDA®/OpenCL™ for increased performance in data-intensive workflows.

Edit Content

Previous slide

Next slide

GPU Design

Over the last three decades, EIZO Rugged Solutions has been providing video and graphics solutions based on our own GPU (graphics processing unit) designs as well as those from other industry leaders such as NVIDIA® and AMD®. We address the evolving needs of the embedded market through cutting-edge innovation, in-house development, and integration of GPU solutions through strategic partnerships. We ensure long life cycles and backward compatibility to support long-term projects and programs through our well managed obsolescence roadmaps. Our products not only provide the high-performance graphics and video required for rugged applications, but they also feature key capabilities such as multiple video formats, ruggedization, built-in codec hardware, real-time raw video capture, and so much more. Our Condor GPGPU products are deployed on numerous high-profile programs, delivering exceptional performance with extended life cycles for mission-critical applications.

Edit Content

Video Format Conversion

As an industry leader in advanced embedded computing engineering, we know what it takes to provide value to our customers. Through our own IP (intellectual property), we provide adaptable solutions that feature video format conversion, video processing, compression, encoding, overlays, and metadata processing. Our video format conversion IP, implemented in FPGAs, facilitates the conversion of any video format to another. As a result, our products can support several video formats such as 3G-SDI, DVI, HDMI, DisplayPort, CoaXPress, and composite video (NTSC, PAL, SECAM, RS-170 one-wire), as well as Sync-on-Green (SOG) RGsB configurations such as STANAG 3350, RS-343, and RS-170 three-wire. You can count on us to accommodate all your I/O needs with our highly adaptable solutions. We can support any video format for both capture and display with our extensive line of products – monitors, video cards, encoders, converter, and splitters.

Edit Content

Previous slide

Next slide

Video Capture

EIZO Rugged Solutions provides video capture solutions (frame grabbers) that interface with a wide variety of sensors and various video formats ranging from legacy analog to modern high resolution digital. Frame-by-frame raw video capture is supported up to the full native refresh rate of the video input source. The captured video pixel data is transferred to host memory over the PCIe bus or directly to GPU memory using GPUDirect® RDMA (NVIDIA®) or DirectGMA (AMD®), minimizing latency between capture and display. Our Condor products offer powerful capabilities and a range of features such as capture, processing, encoding, streaming, and display. Intuitive and easy-to-use software API’s developed by EIZO Rugged Solutions ensure Condor products can be integrated seamlessly into customer configurations. We also provide sample programs to assist with rapid prototyping, testing, and development.

Edit Content

Ruggedization

EIZO Rugged Solutions’ products comply with MIL-STD-810 standards. Ruggedization begins right at the design level, drawing on 30+ years of experience, we carefully select industrial-grade components with advanced PCB and heatsink designs. Our products are designed to work in harsh conditions with extended temperature, shock, and vibration. We also custom-tailor our solutions to meet specific customer requirements. Given that some of our GPU based products consume high power, highly efficient heatsinks are designed to minimize the thermal delta between the GPU to the rails. During production, all products undergo multiple stages of testing before shipping to ensure the product meets the ruggedization requirements. This all-encompassing focus on ruggedization guarantees all products made by EIZO Rugged Solutions can be reliably integrated into a payload and will continue to perform optimally throughout their long life cycle.

We offer three levels of ruggedization – standard, air-cooled, and conduction-cooled. Each air-cooled and conduction-cooled product goes through comprehensive testing to ensure compliance appropriate to that level of ruggedization. Conduction-cooled products come standard with conformal coating. We also offer multiple options for conformal coating materials, including BGA underfills. For complete product customization, our products can be built either using RoHS or leaded build and solder process.

Edit Content

Encoding, Decoding, and Streaming

Compression and decompression of FMV (full motion video) is now a requirement in virtually all ISR deployments. Whether for recording locally without taking up excessive storage space, streaming over limited bandwidth networks for remote monitoring, or for sharing over a local network on a vessel, low latency distribution of FMV is now a common task and is expected to be performed seamlessly without interfering with existing operations.

EIZO Rugged Solutions has been developing and deploying codec streaming products for over 15 years in a wide variety of formats. These include a dedicated H.265/H.264 standalone solution and XMC, 3U VPX or PCIe card level solutions, proprietary codec streaming products, and use of codecs with video capture enabled products that have NVIDIA® GPUs with NVENC/NVDEC. Using these hardware accelerated codecs ensures consistent and low latency throughput with minimal CPU intervention, ensuring precious resources are preserved in the deployment.

Our years of experience and successful deployment of our products to several airborne, naval, and ground platforms, as well as the vast array of products in EIZO Rugged Solutions’ portfolio instills our customers with confidence that their integration of media pipelines at the edge will be successful.

Edit Content

Metadata Insertion/Extraction

While video is often the primary data feed requiring processing, many sensors also provide additional ancillary data along with the video feed. Processing of this data is often critical in the field as it provides information like timestamps, coordinates, angles of the vessel, target frame, and more. Together with video data, this information is used to better interpret critical mission details for increased situational awareness. For SDI sources, this data is typically in the VANC (Vertical Ancillary) space of the video data and encoded in a Key-Length-Value (KLV) format per SMPTE-ST-291M.

Most EIZO Rugged Solutions SDI products (frame grabbers, GPUs, and encoders) support the processing and manipulation of this KLV metadata. We provide sample applications with our frame grabber products, demonstrating this extraction and display of the KLV frame data to the screen. Our encoding products (both box level and GPU based encoders) extract KLV metadata and then synchronously provide this data encapsulated together with the encoded video data in an MPEG2-TS container format. Additionally, our SDI output products allow the insertion of KLV metadata to the SDI outputs with sample applications provided, showing the insertion of sample MISB ST 0601 KLV datasets into the VANC space of the outputs.

Edit Content

Deep Learning

Artificial Intelligence (AI) is here and is becoming more prevalent in the defense and aerospace markets. Advancements in efficient processing of trained neural networks on embedded products have allowed for the deployment of Deep Learning AI technologies into spaces where it was previously infeasible. While training of neural networks is still too intensive and time-consuming to be done in an embedded space, utilization of these models (running inference) is possible at the edge with GPGPU hardware acceleration, allowing for highly advanced and accurate signal processing.

Our solutions based on NVIDIA® Quadro® GPUs enable this functionality through the highly robust CUDA® and CUDA-X AI SDK frameworks. Unification of the NVIDIA driver and support for a wide variety of existing open frameworks allows for flexible development on standard hardware and seamless transition to embedded hardware.

Our latest NVIDIA Quadro Turing™ based products (Condor RTX series) have dedicated Tensor Cores, allowing for AI inferencing to be performed without consuming valuable resources from existing trained models. Furthermore, they can be optimized by using NVIDIA’s own TensorRT™ platform to accelerate workflows further for inference on the NVIDIA Quadro GPUs.

Edit Content

Previous slide

Next slide

GPGPU

In the past, CPUs were the go-to solution when it came down to data processing due to their speed and cost, but now, GPUs have undergone radical transformations that challenge the old status quo. Today with the development of faster and more efficient GPU architectures, GPUs are no longer bound to solely performing image rendering and processing. Many modern advancements offer capabilities for large data set processing and parallel computing which result in vastly increased responsiveness compared to their CPU counterpart.

This shift in the ideology of using the GPU for general processing instead of image processing coined the term GPGPU, or General Processing Graphics Processing Unit. In order to tap into GPU resources, APIs and frameworks were then developed by GPU manufacturers to facilitate this new development including OpenCL™, DirectCompute, AMD® Stream, and NVIDIA® CUDA®. These frameworks have opened the door to specialized areas we now know today, such as computer vision and AI, which are all built on treating GPUs as GPGPUs.

EIZO Rugged Solutions’ graphics products feature real-time processing and low latency in compute-intensive applications. The Condor GPU product line, offered in both NVIDIA and AMD configurations, offer all the benefits of GPGPU out of the box using cutting-edge technology.

Edit Content

CUDA Support

The NVIDIA® CUDA® framework was first introduced in 2007 but has evolved from a simple proprietary GPGPU engine for NVIDIA products to a robust framework offering acceleration of general parallel processing AI, traditional machine-learning, deep learning, and more. CUDA has taken the industry by storm with NVIDIA focusing heavily on support and consistent updates to the framework while also investing in unifying it to work across all NVIDIA GPUs. This allows for flexible upgrade and deployment with little transitional overhead.

The partnership between NVIDIA and EIZO Rugged Solutions has opened the door for CUDA based applications in the embedded rugged space with dedicated NVIDIA based XMC, PCIe, and VPX cards. Highly efficient signal and image processing pipelines are developed, deployed, and reiterated upon, backed by decades of expertise. Traditional environments utilizing FPGAs and other legacy signal processors can now be transitioned to NVIDIA CUDA environments, allowing for a highly simplified upgrade path for both software and hardware components in an integration.

CUDA-X AI™ and NVIDIA’s overall investment in AI also allows the embedded military industry to take advantage of the ability to run highly efficient inferencing on trained neural network models, leading to more accurate image enhancement and classification. Deep learning integration with NVIDIA based Condor GPU products can also be utilized to prolong legacy applications with technologies like NVIDIA NGX and DLSS being utilized to make the once dated camera sensors’ input feeds more usable in the modern era without needing to update underlying hardware/cabling.

Edit Content

NVIDIA Based Products

For the most demanding applications in SWaP-constrained platforms, EIZO Rugged Solutions works closely with NVIDIA^® to provide the most innovative GPU technology for mission-critical applications. As an NVIDIA preferred partner, we have unique access to the highest levels of technical support to help meet the design, production, and product lifecycle requirements of embedded systems. Our XMC, PCIe and 3U VPX video capture and GPGPU cards feature NVIDIA Quadro® GPUs with Pascal™ and Turing™ architecture. These processing powerhouses leverage the latest GPGPU advancements from NVIDIA for machine-learning and artificial intelligence applications as well as standard rendering pipelines.

All NVIDIA based GPGPU and video capture cards offer processing capability supporting multiple rates of TFLOPS, as well as cutting-edge GDDR5 and GDDR6 based memory busses to perform even the most compute-intensive tasks in near real-time. NVIDIA GPUs are an excellent choice for aerospace and defense applications with significant video data processing requirements including video stabilization, image processing, terrain analysis, object tracking, and 3D visualization of data. Support for the CUDA® framework provides the industry’s most advanced and efficient GPGPU framework, opening new doors to the embedded military environment for highly efficient, near real-time, processing of data on platforms.

Edit Content

AMD Based Products

EIZO Rugged Solutions has been an AMD Embedded Partner since 2004, developing high-performance embedded solutions for advanced GPGPU computing to the defense and aerospace markets. The Condor line of video graphics and GPGPU cards feature AMD Radeon™ embedded graphics processors such as the E4690, E8860, and E9171. AMD GPUs are ideal for graphics-intensive avionics applications such as geographic information systems, 360-degree situation awareness, diminished vision enhancement, and more. AMD’s dedicated support for rugged markets has allowed for successful integrations to traditional systems running Windows and Linux platforms, as well as various RTOS systems including VxWorks and INTEGRITY, allowing for flexible GPU integration into safety-critical environments such as cockpit displays. Condor AMD based products offer life cycles of up to 15 years.

AMD embedded solutions are becoming the processor choice for aerospace display and communications systems that demand high-performance graphics interface, advanced security features, floating-point performance, and/or have heavy processing requirements.

GPU Design

Edit Content

Previous slide

Next slide