GPU systems enable AI innovation for critical applications such as autonomous driving, factory floors, video streaming, and cloud gaming with 10x faster data center-to-edge inference and 3x faster training
Super Micro Computer, Inc. (SMCI), a global leader in enterprise computing, storage, networking solutions, and green computing technologies, today announced the expansion of the broadest portfolio of artificial intelligence (AI) GPU servers that integrates The latest NVIDIA Ampere family of GPUs, including the NVIDIA A100, A30 and A2.
The AI performance of Supermicro’s latest NVIDIA-certified systems is ten times higher than previous generations, ensuring AI-enabled applications such as image classification, object detection, reinforcement learning, recommendation, natural language processing (NLP), automatic speech recognition (ASR) ), which can significantly reduce costs and gain deeper insights faster. In addition to inference, Supermicro’s high-performance A100 HGX 8-GPU and 4-GPU servers deliver three times faster AI training and eight times performance in big data analytics compared to previous generations of systems.
“Supermicro continues to lead the GPU market with a broad portfolio of systems optimized for any workload from edge to cloud,” said Charles Liang, President and CEO of Supermicro. “Our total solution for cloud gaming is delivered in a 2U 2-node system. Up to 12 single-width GPUs enable superior density and efficiency. In addition, Supermicro has introduced a new general-purpose GPU platform that brings together all major CPUs, GPUs, and structural and thermal solutions.”
The Supermicro E-403 server is ideal for distributed AI inference applications such as traffic control and office building environments. Supermicro Hyper-E edge servers bring unprecedented inference capabilities to edge applications with up to three A100 GPUs per system. Supermicro now offers complete IT solutions that accelerate collaboration between engineering and design professionals, including NVIDIA-certified servers, storage, network switches, and NVIDIA Enterprise Omniverse software for professional-grade visualization and collaboration.
“Supermicro’s broad portfolio of NVIDIA-certified systems is powered by a complete portfolio of GPUs powered by the NVIDIA Ampere architecture,” said Ian Buck, general manager and vice president of NVIDIA’s Accelerated Computing Group. “Supermicro’s customers benefit from superior performance for all types of modern AI. Workflow — from inference at the edge, to high-performance computing in the cloud, and everything in between.”
Supermicro’s powerful data center 2U and 4U GPU (Redstone, Delta) systems will be first to market, supporting the new Quantum-2 InfiniBand product family and BlueField DPUs. The NVIDIA Quantum-2 InfiniBand solution includes high-bandwidth, ultra-low-latency adapters, switches and cables, as well as complete software that brings high performance to the data center, and runs on the entire Supermicro product line.
Compared to previous generation InfiniBand communication adapters and switches, systems with Quantum-2 InfiniBand offer 400 Gb/s transfer rates, 2X higher bandwidth per port, higher switch density, and 32X higher AI per switch Accelerated and can support both Intel or AMD processors.
As hybrid work environments become the norm, industries need new technologies to ensure technological parity for the workforce. The combination of NVIDIA’s Omniverse Enterprise and Supermicro GPU servers will revolutionize complex 3D workflows, enabling unlimited iteration and faster time-to-market for innovative products. Additionally, NVIDIA’s Omniverse Enterprise, and AI Enterprise on VMware for integrating AI into their enterprise workflows, have been optimized and tested against Supermicro’s NVIDIA-certified systems, enabling geographically dispersed teams to collaborate seamlessly.
The Links: LQ150X1LG81 EPM3512AQC208-10N