GPGPU performance and power estimation using machine learning
Authors:
Top Cited Papers
- 1 February 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15300897,p. 564-576
Abstract
Graphics Processing Units (GPUs) have numerous configuration and design options, including core frequency, number of parallel compute units (CUs), and available memory bandwidth. At many stages of the design process, it is important to estimate how application performance and power are impacted by these options. This paper describes a GPU performance and power estimation model that uses machine learning techniques on measurements from real GPU hardware. The model is trained on a collection of applications that are run at numerous different hardware configurations. From the measured performance and power data, the model learns how applications scale as the GPU's configuration is changed. Hardware performance counter values are then gathered when running a new application on a single GPU configuration. These dynamic counter values are fed into a neural network that predicts which scaling curve from the training data best represents this kernel. This scaling curve is then used to estimate the performance and power of the new application at different GPU configurations. Over an 8× range of the number of CUs, a 3.3× range of core frequencies, and a 2.9× range of memory bandwidth, our model's performance and power estimates are accurate to within 15% and 10% of real hardware, respectively. This is comparable to the accuracy of cycle-level simulators. However, after an initial training phase, our model runs as fast as, or faster than the program running natively on real hardware.Keywords
This publication has 39 references indexed in Scilit:
- PPEP: Online Performance, Power, and Energy Prediction Framework and DVFS Space Exploration2014
- Sources of error in full-system simulation2014
- GPUWattch2013
- Multi2Sim2012
- OpenCL and the 13 dwarfs2012
- A performance analysis framework for identifying potential benefits in GPGPU applications2012
- Break down GPU execution time with an analytical method2012
- CuMAPz2011
- A mechanistic performance model for superscalar out-of-order processorsACM Transactions on Computer Systems, 2009
- MinneSPEC: A New SPEC Benchmark Workload for Simulation-Based Computer Architecture ResearchIEEE Computer Architecture Letters, 2002