Supermicro's multi-node, multi-GPU platform provides high performance for live video streaming, cloud gaming and social networking applications

     Super Micro Computer, a global leader in enterprise computing, storage, networking solutions and green computing technologies,announced details of an innovative new multi-node GPU solution that differs significantly from other existing products on the market. This multi-node system uses shared power and cooling to save up to 10% of the total cost of ownership (TCO). The new 2U 2-node high performance and energy efficient system design allows each node to carry up to 64 cores and 128 PCIe 4.0 channels, with three dual-width PCIe 4.0 Gpus or six single-width PCIe Gpus running at full speed.

     Charles Liang, CEO and President of Supermicro, said: "Our new 2U 2-node multi-GPU server is the ideal platform for live video streaming, advanced cloud gaming and countless social networking applications. With our advanced server Building BlockSolutions® design and high performance architecture, customers can deploy the most energy efficient systems. Supermicro delivers market-leading system resiliency and cost savings to deliver continuous performance to customers. Customers from all over the world have expressed to us a high level of market interest in this unique platform."

     Powered by AMD EPYC 7002 series and next-generation processors, 64 cores and 128 PCIe 4.0 channels per node provide thermal efficiency and streamlined airflow, enabling this high-density multi-GPU system to deliver uninterrupted top performance. This product is ideal for multi-instance, advanced cloud gaming and many other computationally intensive data center applications. The system is equipped with Supermicro's Advanced I/O module (AIOM), which enables fast and resilient network functionality, and can handle large data streams for high-demand AI/ML applications, deep learning training, and inference.

     The unique multi-GPU node design shows excellent maintainability, which is completely different from other existing products on the market. Because of the shared power and heat dissipation resources, the two node drawers in the 2U system can be drawn out separately for easy maintenance. This unique design and the easy accessibility of the GPU reduces maintenance and upgrade costs for high-demand GPU-accelerated applications, such as cloud gaming, which often require high performance and frequent maintenance.