Super Micro Computer, a provider of total IT solutions, announced on the 17th that it participated in the 'International Supercomputing Conference (ISC 2024)' and presented measures to address issues facing the data center and AI/HPC industries.

Supermicro has announced a solution that can help meet the demanding requirements of expanding AI and HPC capacity while reducing data center power consumption.

Supermicro offers complete liquid cooling solutions including cold plates, CDUs, CDMs, and complete cooling towers. Introducing water-cooled servers and infrastructure into a data center can significantly reduce the data center's PUE, reducing total power consumption by up to 40%.

Supermicro 4U 8-GPU liquid-cooled server delivers PetaFlops of AI computing performance in a dense form factor with NVIDIA H100/H200 HGX GPUs.

Future releases include the liquid-cooled X14 8U/6U SuperBlade, rackmount X14 Hyper, and Supermicro X14 BigTwin. A range of HPC-optimized server platforms will support the Intel Xeon 6900 in a compact multi-node form factor with P cores (Performance Cores).

Supermicro continues to lead the industry with the industry’s broadest portfolio of liquid-cooled MGX products, most recently confirming support for Intel’s latest accelerators with the new Intel Gaudi 3 accelerator and AMD MI300X accelerator.

Supermicro SuperBlade, which offers up to 120 nodes per rack, emphasizes that large-scale HPC applications can be run in fewer racks. Supermicro showcased a variety of product lines at ISC 2024, including the Supermicro X14 server with integrated Intel Xeon 6 processors.

Supermicro will showcase and demonstrate a wide range of solutions designed specifically for HPC and AI environments at ISC 2024. Supermicro’s new 4U 8-GPU liquid-cooled server is the flagship product line featuring NVIDIA HGX H100 and H200 GPUs.

It is reported that various servers, including the product in question, will support the NVIDIA B200 HGX GPU as soon as they are released. The new system features advanced GPUs and uses high-speed HBM3 memory to bring more data closer to the GPU than previous systems, accelerating AI training and HPC simulations.

Supermicro SuperCloud Composer for Data Center Management Software was demonstrated, sharing how to monitor and manage an entire data center from a single console, including the status of all liquid-cooled servers.

“Supermicro continues to work with AI and HPC customers to bring the latest technologies to the datacenter, including our comprehensive liquid cooling solutions,” said Charles Liang, President and CEO of Supermicro. “Supermicro’s complete liquid cooling solution can handle up to 100kW per rack, enabling datacenters to reduce TCO (total cost of ownership) while increasing the density of AI and HPC compute.”