top of page

Increasing Embedded Application Performance with GPUs

Embedded graphics enable system developers to boost the performance of a wide range of workloads including medical imaging, image analysis, compute acceleration and artificial intelligence (AI). Leveraging GPUs, embedded graphics can be used to increase application speed and accuracy as well as decrease latency. Many embedded system developers are using embedded graphics solutions in real-world applications such as medical, manufacturing, traffic management. Still, adding a GPU to an embedded system can be a complex task. One approach is to use graphics cards developed for the gaming application segment, which accounts for about a third of all GPUs; however, these cards often do not satisfy key embedded system requirements, such as low system latency, long product availability, and power efficiency. Addressing these issues, ADLINK products greatly simplify the process of adding GPUs to embedded designs. These products can satisfy a wide range of embedded requirements around performance, long life cycle, power consumption and form factor.

Embedded graphics solutions enable system developers, OEMs and systems integrators to significantly improve signal and image processing performance in various application areas including aerospace, maritime, medical, and industrial automation.

Many commercial graphics solutions, such as those developed for gaming applications, have a relatively short lifecycle due to users’ desires for the latest and greatest graphics technology. When embedded solution providers implement these commercial graphics solutions, they may be forced to conduct frequent product certifications, which can be time consuming and expensive. This situation is made more difficult by the relatively few vendors that offer embedded GPU-based computing solutions compared to general-purpose, CPU-based solutions. These are just a few of the factors that may prompt system developers, OEMs and systems integrators to consider integrating embedded GPU-based products from an experienced vendor like ADLINK.


A common theme in embedded application is the need to quickly move external data from sensors and other sources to the GPU for processing. ADLINK achieves this by implementing remote direct memory access (RDMA), a feature of NVIDIA GPUDirect™ technology in NVIDIA® Quadro® GPUs that can boost data throughput by approximately 80 percent (3.6 to 6.5 gigabytes per second). RDMA gives external data sources direct access to the GPU’s external memory. Without this feature, data would be copied into a CPU’s memory before reaching the GPU, which needlessly increases data transmission delay and latency. As an NVIDIA Quadro Embedded Partner with extensive experience in embedded applications, ADLINK is well qualified to provide system developers with solutions using GPUDirect, enabling them to tap into the power of embedded graphics and AI.


With ADLINK’s portfolio of GPU-based solutions and computing platforms, system developers, OEMs, and systems integrators can more easily add embedded graphics to their applications:

ADLINK Embedded Graphics

ADLINK offers two families of embedded graphics products that are an ideal fit for high-resolution, multi-display installations for centralised control and monitoring and high-performance application processing. Mobile PCI Express Modules (MXMs) are well-suited for size, weight and power (SWaP) constrained applications like portable ultrasound, airborne radar and aerial infrared imaging.

PCI Express graphic (PEG) cards connect via a common interface and are easy to integrate and use in many embedded market segments such as healthcare, industrial automation and telecom.


ADLINK Matrix compact fanless embedded computers provide an optimised, highly available computing platform with expandability options including MXM cards. The fanless platforms come in a compact size and support a wide operating temperature range, 5 Grms vibration and cable-free durable structure for use in harsh environments. For applications in less challenging settings, DLAP platforms support performance-hungry workloads (e.g., deep learning) with active cooling for applications requiring very high levels of computing in a limited space.

For solution developers looking for expandable building blocks ADLINK configurable embedded computers can consolidate highly parallel graphics computing, motion control and data acquisition through add-on cards. For applications demanding even greater scalability, ADLINK offers highly configurable motherboards and rackmount industrial chassis that provide an enclosure for ADLINK’s large family of industrial ATX motherboards. The boards feature multiple PCIe/PCI/LAN/USB 3.0, enable immediate multi-tasking deployment and balance performance and expandability.


System developers, OEMs and systems integrators can more easily deploy embedded graphics with help from ADLINK’s large portfolio of computing products that can be designed into a wide range of form factors. Combining its strong expertise serving embedded developers and its close partner relationship with NVIDIA, ADLINK is delivering high-performance, long-life embedded graphics solutions to many market segments.


bottom of page