By Jaan Mannik, Director of Commercial Sales
Edge computing is becoming increasingly important due to the rapid growth of the Internet of Things (IoT), and the need for real-time data processing and analysis. Edge computing involves processing and storing data at the edge of the network, closer to where it is generated, rather than in a centralized location. Conducting these functions closer to where the action takes place means a decision can be made more quickly, producing reduced latency, improved security, greater reliability, and higher performance. This concept is becoming adopted by a wide range of market verticals today, however, edge locations are difficult to optimize and often require reduced physical footprints, non-traditional power & cooling envelopes, and unique rugged design concepts. Enterprise applications requiring the highest end compute for their AI workloads at the Edge are leveraging data-center grade NVIDIA GPUs to get even greater performance. In this blog, I’ll be covering the workhorse advancing HPC at the edge, and how it is being deployed in the real-world AI Transportable applications of today.
The workhorse advancing HPC at the Edge is the GPU (graphics processing unit), by providing a powerful and efficient way to process large amounts of data in real-time. Using GPUs for parallel computing operations is not a new concept. Supercomputers started using GPUs in traditional rack-scale environments back in 2005, when Oak Ridge National Laboratory's Jaguar supercomputer became the first to incorporate GPUs for scientific computing, specifically using GPUs from NVIDIA. The use of GPUs in supercomputers increased in popularity in the following years, due to their ability to perform massive amounts of parallel processing, making them a perfect candidate for high-performance computing applications, such as simulations and modeling, because they can perform many operations simultaneously. As of June 2021, seven of the top ten world's fastest supercomputers use GPUs to accelerate their computations, including the top ranked Fugaku in Japan.
So why is the GPU advancing HPC at the Edge? A single NVIDIA A100 80GB GPU can greatly increase the performance, compared to a CPU, and requires a much smaller footprint. As NVIDIA explains “On state-of-the-art conversational AI models like BERT, A100 accelerates inference throughput up to 249x over CPU.” This makes GPUs ideal for high-performance computing at the Edge, where fast processing is essential to ensure timely decision-making and response. A single 3U rugged server supporting 4 NVIDIA A100 GPUs can replace an entire rack of CPU-based servers traditionally found clustered in a datacenter, which is a game changer for Edge applications where size, space, power, and cooling are problematic. For example, in autonomous vehicles, GPUs can process sensor data and make real-time decisions to control the vehicle. In agriculture, GPUs can identify weeds to eradicate to help crops grow without using pesticides. In healthcare, GPUs can analyze medical images in real-time to provide more accurate diagnoses. One Stop Systems 3U SDS A100 GPU Server is an NVIDIA Certified rugged HPC platform being deployed in these environments today, supporting up to four A100 GPUs and 500TB of NVMe flash storage, in a 3U compact chassis, optimized for the very edge.
Overall, edge computing can help organizations improve their operational efficiency, reduce costs, and improve the performance and reliability of their applications. Adding enterprise-class GPUs brings the power of the data center to the very Edge, by providing a powerful and efficient way to process large amounts of data in real-time where it is most needed. This ‘workhorse’ is an incredibly important technology that is pushing the limits of Edge computing, enabling new and innovative applications that were not possible before.
Click the buttons below to share this blog post!
The rugged edge computing landscape is becoming increasingly complex with new generations of technologies, such as the latest AI focused GPUs, releasing annually rather than every 2-3 years. Whether the end application is commercial or defense, rugged edge servers must not only deliver cutting-edge compute performance but also withstand extreme environmental conditions.
When the PCI-SIG formally added support for 675W add-in card devices in the PCI Express Card Electromechanical (CEM) specification in August 2023, NVIDIA’s most powerful CEM GPU, the NVIDIA H100 80GB had a maximum power consumption of 350W. While some devices were starting to push the limits of datacenter thermodynamics – high density systems of many 675W devices seemed like a distant reality. However, with power constraints uncapped and the need for higher performing GPUs skyrocketing, the industry quickly came out with devices taking full advantage of the new specification capability. NVIDIA quickly replaced the H100 80GB with the H100 NVL, increasing power density to 400W. While this small jump was manageable for existing installations, NVIDIA then dove all-in with the H200 NVL released in late 2024 at 600W. The rapid transition from 350W to 600W has put power and cooling technologies in the spotlight in a race to solve this next generation challenge.
The advent of technology has always brought about significant changes to various industries, and the transportation sector is no exception. Among the most transformative innovations in recent years is the development of autonomous vehicles, particularly trucks. The potential for autonomous trucks to revolutionize freight transport is immense, raising the fundamental question: will these technological advancements make human drivers obsolete? To explore this question, we must consider the current state of autonomous driving technology, the economic implications, and the societal impact of removing human drivers from the equation.
Get the latest news, product releases, upcoming events, blog posts and more!
Rein Mannik
June 02, 2023
Impressive and convincing the need for having information available so that decisions can be earlier. Almost like integrating some analogue computing with digital computing.