Open Compute gets liquid cooling

Penguin HPC systems will use Asetek cooling.

Penguin Computing announced that it has selected Asetek to be the provider of the liquid cooling solution for its Open Compute-based high performance computing solution. Penguin will utilize the Asetek RackCDU cooling system as a standard component of its Tundra Extreme Scale HPC server product line.


The first order is for 21 of the rackmount systems, which can support more than 1200 compute nodes with their direct-to-chip cooling loop. Asetek expects the total number of RackCDU systems for Penguin's off-the-rack HPC solution to exceed 300 units over the next three years.

These systems will likely be used to fulfill Penguin's contract for what will be one of the largest Open Compute-based HOPC cluster solutions with the Lawrence Livermore National Laboratory. The Tundra HPC system will be installed to meet the requirements of the National Nuclear Security Administration's Commodity Technology System (CTS-1) which will be installed at the research facilities at Lawrence Livermore, Sandia, and Los Alamos research facilities.

In previous testing with the Asetek RackCDU system, Lawrence Livermore was able to verify the cooling and density increase claims of Asetek when compared to standard non-liquid cooling methods.