GIGABYTE Announces New Liquid Cooled Solutions for NVIDIA HGX H200
G593 Series – Tailored Cooling
The GPU-centric G593 series is custom engineered to house an 8-GPU baseboard, and its design had foresight for both air and liquid cooling. The compact 5U chassis leads the industry in its readily scalable nature, fitting up to sixty-four GPUs in a single rack and supporting 100kW of IT hardware. This helps to consolidate the IT hardware, and in turn, decrease the data center footprint. The G593 series servers for DLC are in response to the rising customer demand for greater energy efficiency. Liquids have a higher thermal conductivity than air, so they can rapidly and effectively remove heat from hot components to maintain lower operating temperatures. And by relying on water and heat exchangers, the overall energy consumption of the data center is reduced.
“We offer a great AI scaling GIGABYTE solution using NVIDIA HGX H200 GPU,” said Vincent Wang, VP of Sales at Giga Computing. “Due to the complexity of enterprise data centers, there is a need to ensure the infrastructure can handle the computational demand and complexity of AI, ML, and data science models. This increasing complexity necessitates greater optimization. We can develop and invest in scalable AI infra. And we can cover all aspects of AI data center infra services from deployment to software stack in cooperation with the NVIDIA NVAIE platform.”
GIGABYTE has now released both air-cooled and DLC versions of its G593 series for both platforms, NVIDIA HGX H200 and NVIDIA HGX H100. Forthcoming newly designed GIGABYTE servers for NVIDIA HGX B200A will also have the option of air or liquid cooling. For rack-scale deployment of all these mentioned NVIDIA HGX platforms, GIGABYTE has already introduced GIGAPOD as its answer to the need for a complete supercluster with 256x NVIDIA H100 GPUs. For DLC servers, this solution is composed of five racks with four of them populated with eight G593 servers each. And for air cooling, a nine-rack solution can house the same thirty-two G593 servers. Excellent interconnectivity is achieved using NVIDIA NVLink and NVIDIA NVSwitch, and systems are also paired with InfiniBand to interconnect between nodes in the cluster. All in all, a complete cluster can easily tackle large-scale model training, scientific simulations, and more.
- G593-ZD1-LAX3: AMD EPYC 9004 processors
- G593-SD1-LAX3: 5th Gen Intel Xeon processors