One of the World’s Largest Open Compute Based Installations
Reduce power consumption and the associated costs at data centers worldwide. Create a model to be leveraged by other HPC sites to achieve improved energy efficiency.
Solution:
Asetek® Liquid Cooled Penguin® Computing Tundra™ Extreme Scale HPC Clusters. In 2017, more than 125 RackCDUs and over 6,900 compute nodes were installed.
In 2018, Penguin Computing is installing Asetek’s RackCDU Direct-to-Chip (D2C) liquid cooling solution and D2C coolers and loops for OCP-based Tundra Relion™ nodes using both NVIDIA®’s P100 GPU accelerators and Intel® Xeon® ES-2600 processors. This installation is expected to include 14 RackCDUs and 860 nodes.
Installation Highlights:
The Tri-lab effort has reduced costs, increased operational efficiencies, & facilitated collaborations that benefit U.S. security, support academia, & advanced the technology that promotes economic competitiveness.
Additional Information: Press Release
TOP 500 Ranking - June 2018: LANL Grizzly (#136), SNL Serrano (#168), SNL Cayenne (#169), SNL Eclipse (191) and LLNL Topaz (#449)
Green500 Ranking - June 2018: SNL Serrano (#75), SNL Cayenne (#76), LANL Grizzly (#77), SNL Eclipse (#86) and LLNL Topaz (#87)
