ePrints@IIScePrints@IISc Home | About | Browse | Latest Additions | Advanced Search | Contact | Help

Scheduling Page Table Walks for Irregular GPU Applications

Shin, Seunghee and Cox, Guilherme and Oskin, Mark and Loh, Gabriel H and Solihin, Yan and Bhattacharjee, Abhishek and Basu, Arkaprava (2018) Scheduling Page Table Walks for Irregular GPU Applications. In: 45th ACM/IEEE Annual International Symposium on Computer Architecture (ISCA), JUN 01-06, 2018, Los Angeles, CA, pp. 180-192.

[img] PDF
45th _ACM_IEEE_Ann_Int_Sym_Com_Arc_180-192_2018.pdf - Published Version
Restricted to Registered users only

Download (1MB) | Request a copy
Official URL: https://doi.org/10.1109/ISCA.2018.00025

Abstract

Recent studies on commercial hardware demonstrated that irregular GPU applications can bottleneck on virtual-to-physical address translations. In this work, we explore ways to reduce address translation overheads for such applications. We discover that the order of servicing a GPU's address translation requests (specifically, page table walks) plays a key role in determining the amount of translation overhead experienced by an application. We find that different SIMD instructions executed by an application require vastly different amounts of work to service their address translation needs, primarily depending upon the number of distinct pages they access. We show that better forward progress is achieved by prioritizing translation requests from the instructions that require less work to service their address translation needs. Further, in the GPU's Single-Instruction-Multiple-Thread (SIMT) execution paradigm, all threads that execute in lockstep (wavefront) need to finish operating on their respective data elements (and thus, finish their address translations) before the execution moves ahead. Thus, batching walk requests originating from the same SIMD instruction could reduce unnecessary stalls. We demonstrate that the reordering of translation requests based on the above principles improves the performance of several irregular GPU applications by 30% on average.

Item Type: Conference Proceedings
Additional Information: 45th ACM/IEEE Annual International Symposium on Computer Architecture (ISCA), Los Angeles, CA, JUN 01-06, 2018
Keywords: Computer architecture; GPU; Virtual address
Department/Centre: Division of Electrical Sciences > Computer Science & Automation
Depositing User: Id for Latest eprints
Date Deposited: 15 Mar 2019 05:29
Last Modified: 15 Mar 2019 05:29
URI: http://eprints.iisc.ac.in/id/eprint/61961

Actions (login required)

View Item View Item