Improving GPGPU Energy-Efficiency through Concurrent Kernel Execution and DVFS

Current generation GPUs can accelerate high-performance, compute-intensive applications by exploiting massive thread-level paral-lelism. The high performance, however, comes at the cost of in-creased power consumption. Recently, commercial GPGPU archi-tectures have introduced support for concurrent kernel execution to better utilize the computational/memory resources and thereby improve overall throughput. In this paper, we argue and experimen-tally validate the benefits of concurrent kernels towards energy-efficient execution. We design power-performance models to care-fully select the appropriate kernel combinations to be executed concurrently, the relative contributions of the kernels to the thread mix, along with the frequency choices for the cores and the mem-ory to achieve high performance per watt metric. Our experimental evaluation shows that the concurrent kernel execution in combi-nation with DVFS can improve energy-efficiency by up to 34.5% compared to the most energy-efficient sequential execution. 1