Please use this identifier to cite or link to this item:
For citation please use:
Main Title: SLC: Memory Access Granularity Aware Selective Lossy Compression for GPUs
Author(s): Lal, Sohan
Lucas, Jan
Juurlink, Ben
Type: Conference Object
Is Part Of: 10.14279/depositonce-9156
Language Code: en
Abstract: Memory compression is a promising approach for reducing memory bandwidth requirements and increasing performance, however, memory compression techniques often result in a low effective compression ratio due to large memory access granularity (MAG) exhibited by GPUs. Our analysis of the distribution of compressed blocks shows that a significant percentage of blocks are compressed to a size that is only a few bytes above a multiple of MAG, but a whole burst is fetched from memory. These few extra bytes significantly reduce the compression ratio and the performance gain that otherwise could result from a higher raw compression ratio. To increase the effective compression ratio, we propose a novel MAG aware Selective Lossy Compression (SLC) technique for GPUs. The key idea of SLC is that when lossless compression yields a compressed size with few bytes above a multiple of MAG, we approximate these extra bytes such that the compressed size is a multiple of MAG. This way, SLC mostly retains the quality of a lossless compression and occasionally trades small accuracy for higher performance. We show a speedup of up to 35% normalized to a state-of-the-art lossless compression technique with a low loss in accuracy. Furthermore, average energy consumption and energy-delay- product are reduced by 8.3% and 17.5%, respectively.
Issue Date: 16-May-2019
Date Available: 11-Feb-2019
DDC Class: 004 Datenverarbeitung; Informatik
Subject(s): GPU
memory access granularity
Selective Lossy Compression
Sponsor/Funder: EC/H2020/688759/EU/Low-Power Parallel Computing on GPUs 2/LPGPU2
Proceedings Title: 2019 Design, Automation & Test in Europe Conference & Exhibition (DATE)
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Publisher Place: New York, NY
Publisher DOI: 10.23919/DATE.2019.8714810
EISSN: 1558-1101
ISBN: 978-3-9819263-2-3
Appears in Collections:FG Architektur eingebetteter Systeme » Publications

Files in This Item:

Accepted manuscript

Format: Adobe PDF | Size: 252.46 kB
DownloadShow Preview

Version History
Version Item Date Summary
2 10.14279/depositonce-8185.2 2019-11-21 09:38:32.317 add publisher metadata
1 10.14279/depositonce-8185 2019-02-11 13:27:59.0
Item Export Bar

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