As innovation continues to progress, so does high-performance computer (HPC) processing power.
Because of that, HPC clusters require more support than ever before. High performance GPUs and CPUs are being combined into denser configurations that lead to higher wattage densities at both the node and rack level. The days of rack power requirements below 20kW are long gone, with modern rack densities easily topping 80kW and beyond.
Data centers are facing new thermal management challenges as traditional air cooling is unable to keep up with increased demand. When you’re running clusters dedicated to critical computing at max capacity for 24 hours a day, 7 days a week, better solutions are needed.
In data centers, liquid cooling offers the increased efficiency and reduced costs that operations managers are seeking. By lowering processor temperatures through liquid cooling, operators can realize a 4% reduction in solution times and minimize latency by maximizing cluster interconnect density.