Please use this identifier to cite or link to this item:
Main Title: Local memory-aware kernel perforation
Author(s): Maier, Daniel
Cosenza, Biagio
Juurlink, Ben
Type: Conference Object
Language Code: en
Abstract: Many applications provide inherent resilience to some amount of error and can potentially trade accuracy for performance by using approximate computing. Applications running on GPUs often use local memory to minimize the number of global memory accesses and to speed up execution. Local memory can also be very useful to improve the way approximate computation is performed, e.g., by improving the quality of approximation with data reconstruction techniques. This paper introduces local memory-aware perforation techniques specifically designed for the acceleration and approximation of GPU kernels. We propose a local memory-aware kernel perforation technique that first skips the loading of parts of the input data from global memory, and later uses reconstruction techniques on local memory to reach higher accuracy while having performance similar to state-of-the-art techniques. Experiments show that our approach is able to accelerate the execution of a variety of applications from 1.6× to 3× while introducing an average error of 6%, which is much smaller than that of other approaches. Results further show how much the error depends on the input data and application scenario, the impact of local memory tuning and different parameter configurations.
Issue Date: 2018
Date Available: 4-Jun-2018
DDC Class: 004 Datenverarbeitung; Informatik
Subject(s): approximate computing
kernel perforation
Proceedings Title: Proceedings of 2018 IEEE/ACM International Symposium on Code Generation and Optimization (CGO’18)
Publisher: Association for Computing Machinery (ACM)
Publisher Place: New York, NY, USA
Publisher DOI: 10.1145/3168814
Page Start: 278
Page End: 287
ISBN: 978-1-4503-5617-6
Appears in Collections:FG Architektur eingebetteter Systeme » Publications

Files in This Item:
File Description SizeFormat 
MaierCGO18.pdf688.83 kBAdobe PDFThumbnail

Items in DepositOnce are protected by copyright, with all rights reserved, unless otherwise indicated.