کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
460345 696326 2016 14 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Auto-tuning for GPGPU applications using performance and energy model
موضوعات مرتبط
مهندسی و علوم پایه مهندسی کامپیوتر شبکه های کامپیوتری و ارتباطات
پیش نمایش صفحه اول مقاله
Auto-tuning for GPGPU applications using performance and energy model
چکیده انگلیسی


• Performance and Energy models are proposed for GPGPU software applications with high fidelity, which can be used for tuning the software parameters.
• Auto-tuning framework based on simulated annealing and genetic algorithm is proposed for GPGPU software based on the above performance and energy models such that software can be tuned based on user requirements.
• This work is the first that combines performance and energy tuning automatically for GPGPU software.

The general-purpose graphic processing unit (GPGPU) is a popular accelerator for general applications such as scientific computing because the applications are massively parallel and the significant power of parallel computing inheriting from GPUs. However, distributing workload among the large number of cores as the execution configuration in a GPGPU is currently still a manual trial-and-error process. Programmers try out manually some configurations and might settle for a sub-optimal one leading to poor performance and/or high power consumption. This paper presents an auto-tuning approach for GPGPU applications with the performance and power models. First, a model-based analytic approach for estimating performance and power consumption of kernels is proposed. Second, an auto-tuning framework is proposed for automatically obtaining a near-optimal configuration for a kernel computation. In this work, we formulated that automatically finding an optimal configuration as the constraint optimization and solved it using either simulated annealing (SA) or genetic algorithm (GA). Experiment results show that the fidelity of the proposed models for performance and energy consumption are 0.86 and 0.89, respectively. Further, the optimization algorithms result in a normalized optimality offset of 0.94% and 0.79% for SA and GA, respectively.

Figure optionsDownload as PowerPoint slide

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Journal of Systems Architecture - Volume 62, January 2016, Pages 40–53
نویسندگان
, , ,