ApproxTuner: A Compiler and Runtime System for Adaptive Approximations
Manually optimizing the tradeoffs between accuracy, performance and energy for resource-intensive applications with flexible accuracy or precision requirements is extremely difficult. We present ApproxTuner, an automatic framework for accuracy-aware optimization of tensor-based applications while requiring only high-level end-to-end quality specifications. ApproxTuner implements and manages approximations in algorithms, system software, and hardware.
The key contribution in ApproxTuner is a novel three-phase approach to approximation-tuning that consists of development-time, install-time, and run-time phases. Our approach decouples tuning of hardware-independent and hardware-specific approximations, thus providing retargetability across devices. To enable efficient autotuning of approximation choices, we present a novel accuracy-aware tuning technique called predictive approximation-tuning, which speeds up autotuning by analytically predicting the accuracy impacts of approximations.
We evaluate ApproxTuner across 10 convolutional neural networks (CNNs) and a combined CNN and image processing benchmark. For the evaluated CNNs, using only hardware-independent approximation choices we achieve a mean speedup of 2.1x (max 2.7x) on a GPU, and 1.3x mean speedup (max 1.9x) on the CPU, while staying within 1 percentage point of inference accuracy loss. For two different accuracy-prediction models, ApproxTuner speeds up tuning by 12.8x and 20.4x compared to conventional empirical tuning while achieving comparable benefits.
Tue 2 MarDisplayed time zone: Eastern Time (US & Canada) change
11:10 - 12:10
|GPTune: Multitask Learning for Autotuning Exascale Applications|
Yang Liu , Wissam M. Sid-Lakhdar Lawrence Berkeley National Laboratory, Osni Marques Lawrence Berkeley National Laboratory, Xinran Zhu Cornell University, Chang Meng Emory University, James W. Demmel UC Berkeley, Xiaoye S. Li Lawrence Berkeley National LaboratoryLink to publication
|I/O Lower Bounds for Auto-tuning of Convolutions in CNNs|
Xiaoyang Zhang Institute of Computing Technology, Chinese Academy of Sciences, Junmin Xiao Institute of Computing Technology, Chinese Academy of Sciences, Guangming Tan Institute of Computing Technology, Chinese Academy of SciencesLink to publication
|ApproxTuner: A Compiler and Runtime System for Adaptive Approximations|
Hashim Sharif University of Illinois at Urbana Champaign, Yifan Zhao University of Illinois at Urbana Champaign, Maria Kotsifakou Runtime Verification, Inc., Akash Kothari University of Illinois at Urbana Champaign, Ben Schreiber University of Illinois at Urbana Champaign, Elizabeth Wang University of Illinois at Urbana Champaign, Yasmin Sarita Cornell University, Nathan Zhao University of Illinois at Urbana-Champaign, Keyur Joshi University of Illinois at Urbana-Champaign, Vikram S. Adve University of Illinois at Urbana-Champaign, Sasa Misailovic University of Illinois at Urbana-Champaign, Sarita Adve University of Illinois at Urbana-ChampaignLink to publication