Quantifying the effectiveness of load balance algorithms

Load balance is critical for performance in large parallel applica-tions. An imbalance on today’s fastest supercomputers can force hundreds of thousands of cores to idle, and on future exascale ma-chines this cost will increase by over a factor of a thousand. Im-proving load balance requires a detailed understanding of the amount of computational load per process and an application’s simulated domain, but no existing metrics sufficiently account for both fac-tors. Current load balance mechanisms are often integrated into applications and make implicit assumptions about the load. Some strategies place the burden of providing accurate load information, including the decision on when to balance, on the application. Ex-isting application-independent mechanisms simply measure the ap-plication load without any knowledge of application elements, which limits them to identifying imbalance without correcting it