Supermicro, Inc., a Total IT Solution Provider for AI/ML, HPC, Cloud, Storage, and 5G/Edge, is introducing a wide range of new systems which are fully optimized for edge and embedded workloads. Several of these new compact servers, which are based on the latest Intel Xeon 6 SoC processor family (formerly codenamed Granite Rapids-D), empower businesses to optimize real-time AI inferencing and enable smarter applications across many key industries.
“As the demand for Edge AI solutions grows, businesses need highly reliable, compact systems that can process data at the edge in real-time,” said Charles Liang, president and CEO of Supermicro. “At Supermicro, we design and deploy the industry’s broadest range of application optimized systems from the data center to the far edge. Our latest generation of edge servers deliver advanced AI capabilities for enhanced efficiency and decision-making close to where the data is generated. With up to 2.5 times core count increase at the edge with improved performance per watt and per core, these new Supermicro compact systems are fully optimized for workloads such as Edge AI, telecom, networking, and CDN.”
Supermicro’s new SYS-112D series systems are designed to run high-performance Edge AI solutions and feature the recently launched Intel Xeon 6 SoC with P-cores. These servers feature increased performance, improved performance per watt, and higher memory bandwidth compared to previous generations of systems. In addition, the new servers include AI acceleration, Intel QuickAssist Technology with wireless protocols, Intel vRAN Boost Technology, Intel Data Streaming Accelerator, and more.
Supermicro’s SYS-112D-36C-FN3P features the Intel Xeon 6 SoC with 36 P-cores, dual 100 GbE QSFP28 ports, up to 512GB of DDR5 memory, and one PCIe 5.0 FHFL slot for a GPU or other add-on card. Combined with Intel’s onboard Media Acceleration and QuickAssist technologies, this makes the system ideal for Edge AI and media workloads, and with a chassis only 399mm/15.7inches deep and with front I/O access, it can easily be deployed in space-constrained environments or embedded in larger systems. Another server based on the same platform, the SYS-112D-42C-FN8P, provides a more telco-optimized configuration, featuring 8 25GbE ports, built-in GNSS and time sync technology, and an Intel Xeon 6 SoC model featuring Intel vRAN Boost. The combination of these features makes this model an all-in-one platform for various workloads in the RAN network.
Supermicro is also introducing two new compact systems, the SYS-E201-14AR and SYS-E300-14AR, which are optimized for IoT and AI inferencing at the far edge. Both systems feature the 15th Gen Intel Core Ultra processors (codenamed Arrow Lake), which features up to 24 cores and an onboard NPU (Neural Processing Unit) AI accelerator. Both systems have two 2.5 GbE network ports, and connectors for HDMI, Display, and USB, and are optimized for enterprise edge use cases. The SYS-E300 can also be expanded to feature a single PCIe 5.0 x16 slot, allowing for the installation of a PCIe GPU card, enabling the system to expand its performance for Edge AI applications in security & surveillance, retail, healthcare, manufacturing and more.
In the edge data center, Supermicro’s edge AI systems can now be installed with the recently launched Intel Xeon 6700/6500 series processor with P-cores. This processor group is designed for the enterprise data center, aiming for a strong balance between performance and efficiency and delivering an average 1.4x better performance than the previous generation across a wide range of enterprise workloads. Supermicro’s 2U Edge AI product family, such as the SYS-212B-FLN2T, combines Intel’s new processor with up to 6 single-width GPU accelerators in a short-depth, front I/O form factor that can be deployed at the enterprise edge as well as in telco and space-constrained environments.
