Using Software to Make Datacenters Greener
It’s no secret that supercomputers require a lot of energy to run. Roadrunner was decommissioned because it simply used too much energy. TH1, the fastest supercomputer in the world from 2010-11, uses over four megawatts of power, which is the same amount of electricity needed to power approximately 3,500 homes. Organizations have tried using different cooling methods to keep the cost of running a datacenter down, while Adaptive Computing is going about things a different way, with workload management software.
As the future of green computing continues to move forward, workload management will get more and more attention. Workload management software is middleware that acts like a chorus instructor, determining which portions of a datacenter will be used at any given time. The portions that aren’t being used are powered down, saving energy and making datacenters more efficient. Workload management software accomplishes this without sacrificing performance.
In other words, workload management software sort of makes datacenters like sharks. If a shark stops moving, it will drown. So, to get sleep, it will shut off part of its brain to save energy and stay alive.
Rob Clyde is the CEO of Adaptive Computing, which produces workload management software to make supercomputers and company’s private clouds more efficient. Adaptive’s workload management software is currently being used on the Beacon supercomputer.
“Today’s Beacon supercomputer is the top green system in the world,” Clyde said. “It has the best gigaflops per watt rating at 2.5 gigaflops per watt, which is six times more efficient than Roadrunner.”
According to Clyde, there are three ways to make supercomputers more efficient.
The average server is 8.5 percent in datacenters. Using workload management software, servers can see increased utilization by consolidating workloads, increasing energy savings by three to four times. Datacenters can go a step further by scheduling workloads so there are instances when the datacenter is not in use. The third way is to use more efficient processors. By selecting the appropriate chip for the job, energy can be saved.
Clyde mentioned that a typical datacenter could become more effective while cutting their energy costs almost in half with workload management software.