Supermicro Accelerates Everything at ISC23
Both AI and HPC depend on cutting-edge computing hardware and software to accomplish their respective goals. HPC aims to simulate physical phenomena , while AI requires large quantities of data and robust analytical tools to learn from that data. Creating HPC environments that are adaptable and cost-effective enough to support a company’s current and long-term AI strategies can be difficult, as the choices made around the system and GPU architecture will greatly affect the final performance.
As a leading supplier of AI infrastructure, Supermicro’s turn-key designs leverage its vast experience in building some of the world’s largest AI clusters
Liquid Cooled AI Platforms
In a fully populated rack at ISC23, Supermicro showcased the industry’s first NVIDIA HGX H100 8 and 4-way GPU server with liquid cooling, containing dual Intel or AMD CPUs and with interconnected NVIDIA HGX H100 Tensor Core GPUs. Using liquid cooling reduces the power consumption of data centers by up to 40%, resulting in lower operating costs. Both systems provide up to 30x performance and efficiency in today’s large transformer models with faster GPU-GPU interconnect speed and PCIe 5.0 based networking and storage. You can find out more from the press release here.
Also on show was the new Liquid Cooled AI development platform from Supermicro and NVIDIA, which is changing how AI developers advance the rapidly evolving AI-enabled world. This new solution, including an operating system and NVIDIA AI Enterprise software, is ready to go right out of the box, featuring dual 4th Gen Intel® Xeon® processors and four NVIDIA A100 GPUs. Perfect for small & medium businesses or enterprise customers looking to use AI as a competitive advantage. Download the solution brief here.
Supermicro and Intel-powered accelerated AI Building blocks.
Intel’s Jeff McVeigh delivered an enlightening keynote presentation at ISC23, highlighting how AI-accelerated HPC has the potential to revolutionize the way we tackle the world’s most challenging problems. Supermicro’s Ray Pang joined the keynote to discuss how ground-breaking technology like Supermicro’s new 8U PVC system, which accommodates 8xOAM Intel® Data Series GPUs and dual 4th Gen Intel® Xeon® Scalable Processors, has been specifically engineered to cater to large-scale AI training, generative AI, and HPC. You can watch the keynote in full here.
Maximizing your storage performance
Addressing the crowds at ISC23 alongside our storage partner Weka, Supermicro also showcased its broad range of WekaIO-ready systems. These high-performance shared, and parallel file storage solutions are optimized for NVMe flash and designed for both speed and scalability – critical factors in any HPC and AI strategy. Find out more here.
With a comprehensive range of Servers and Solutions specifically designed to support AI workloads, as showcased at ISC23, Supermicro is ready to help you accelerate everything.