Energy, Power, and Performance Characterization of GPGPU Benchmark Programs
Authors:
- Jared CoplinJCJared Coplin
- Martin BurtscherMBMartin Burtscher
- 1 May 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1190-1199
Abstract
This paper studies the effects on energy consumption, power draw, and runtime of a modern compute GPU when changing the core and memory clock frequencies, enabling or disabling ECC, using alternate implementations, and varying the program inputs. We evaluate 34 applications from 5 benchmark suites and measure their power draw over time on a K20c GPU. Our results show that changing the frequency or the program implementation can alter the energy, power, and performance by a factor of two or more. Interestingly, some changes affect these three aspects very unevenly. ECC can greatly increase the runtime and energy consumption, but only on memory-bound codes. Compute-bound codes tend to behave quite differently from memory-bound codes, in particular regarding their power draw. On irregular programs, a small change in frequency can result in a large change in runtime and energy consumption.Keywords
This publication has 21 references indexed in Scilit:
- Empirical characterization of power efficiency for large scale data processing2015
- Evaluating the performance and energy efficiency of n-body codes on multi-core CPUs and GPUs2013
- Effects of Dynamic Voltage and Frequency Scaling on a K20 GPU2013
- Energy Analysis of Parallel Scientific Kernels on Multiple GPUs2012
- Understanding Power Measurement Implications in the Green500 List2010
- Designing Power-Aware Collective Communication Algorithms for InfiniBand Clusters2010
- Towards optimizing energy costs of algorithms for shared memory architectures2010
- Optimistic parallelism requires abstractions2007
- Exploring the Energy-Time Tradeoff in MPI Programs on a Power-Scalable Cluster2005
- Studying Thermal Management for Graphics-Processor Architectures2005