Parallel GPU architecture simulation framework exploiting work allocation unit parallelism

Sangpil Lee, Won Woo Ro

Research output: Chapter in Book/Report/Conference proceedingConference contribution

17 Citations (Scopus)

Abstract

GPU computing is at the forefront of high-performance computing, and it has greatly affected current studies on parallel software and hardware design because of its massively parallel architecture. Therefore, numerous studies have focused on the utilization of GPUs in various fields. However, studies of GPU architectures are constrained by the lack of a suitable GPU simulator. Previously proposed GPU simulators do not have sufficient simulation speed for advanced software and architecture studies. In this paper, we propose a new parallel simulation framework and a parallel simulation technique called work-group parallel simulation in order to improve the simulation speed for modern many-core GPUs. The proposed framework divides the GPU architecture into parallel and shared components, and it determines which GPU component can be effectively parallelized and can work correctly in multithreaded simulation. In addition, the work-group parallel simulation technique effectively boosts the performance of parallelized GPU simulation by eliminating the synchronization overhead. Experimental results obtained using a simulator with the proposed framework show that the proposed parallel simulation technique has a speed-up of up to 4.15 as compared to an existing sequential GPU simulator on an 8-core machine providing minimized cycle errors.

Original languageEnglish
Title of host publicationISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software
Pages107-117
Number of pages11
DOIs
Publication statusPublished - 2013 Aug 19
Event2013 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2013 - Austin, TX, United States
Duration: 2013 Apr 212013 Apr 23

Other

Other2013 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2013
CountryUnited States
CityAustin, TX
Period13/4/2113/4/23

Fingerprint

Simulators
Parallel architectures
Graphics processing unit
Computer hardware
Synchronization

All Science Journal Classification (ASJC) codes

  • Software

Cite this

Lee, S., & Ro, W. W. (2013). Parallel GPU architecture simulation framework exploiting work allocation unit parallelism. In ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software (pp. 107-117). [6557151] https://doi.org/10.1109/ISPASS.2013.6557151
Lee, Sangpil ; Ro, Won Woo. / Parallel GPU architecture simulation framework exploiting work allocation unit parallelism. ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software. 2013. pp. 107-117
@inproceedings{cbd5502a4f794b38a3001324125c3c27,
title = "Parallel GPU architecture simulation framework exploiting work allocation unit parallelism",
abstract = "GPU computing is at the forefront of high-performance computing, and it has greatly affected current studies on parallel software and hardware design because of its massively parallel architecture. Therefore, numerous studies have focused on the utilization of GPUs in various fields. However, studies of GPU architectures are constrained by the lack of a suitable GPU simulator. Previously proposed GPU simulators do not have sufficient simulation speed for advanced software and architecture studies. In this paper, we propose a new parallel simulation framework and a parallel simulation technique called work-group parallel simulation in order to improve the simulation speed for modern many-core GPUs. The proposed framework divides the GPU architecture into parallel and shared components, and it determines which GPU component can be effectively parallelized and can work correctly in multithreaded simulation. In addition, the work-group parallel simulation technique effectively boosts the performance of parallelized GPU simulation by eliminating the synchronization overhead. Experimental results obtained using a simulator with the proposed framework show that the proposed parallel simulation technique has a speed-up of up to 4.15 as compared to an existing sequential GPU simulator on an 8-core machine providing minimized cycle errors.",
author = "Sangpil Lee and Ro, {Won Woo}",
year = "2013",
month = "8",
day = "19",
doi = "10.1109/ISPASS.2013.6557151",
language = "English",
isbn = "9781467357777",
pages = "107--117",
booktitle = "ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software",

}

Lee, S & Ro, WW 2013, Parallel GPU architecture simulation framework exploiting work allocation unit parallelism. in ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software., 6557151, pp. 107-117, 2013 IEEE International Symposium on Performance Analysis of Systems and Software, ISPASS 2013, Austin, TX, United States, 13/4/21. https://doi.org/10.1109/ISPASS.2013.6557151

Parallel GPU architecture simulation framework exploiting work allocation unit parallelism. / Lee, Sangpil; Ro, Won Woo.

ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software. 2013. p. 107-117 6557151.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Parallel GPU architecture simulation framework exploiting work allocation unit parallelism

AU - Lee, Sangpil

AU - Ro, Won Woo

PY - 2013/8/19

Y1 - 2013/8/19

N2 - GPU computing is at the forefront of high-performance computing, and it has greatly affected current studies on parallel software and hardware design because of its massively parallel architecture. Therefore, numerous studies have focused on the utilization of GPUs in various fields. However, studies of GPU architectures are constrained by the lack of a suitable GPU simulator. Previously proposed GPU simulators do not have sufficient simulation speed for advanced software and architecture studies. In this paper, we propose a new parallel simulation framework and a parallel simulation technique called work-group parallel simulation in order to improve the simulation speed for modern many-core GPUs. The proposed framework divides the GPU architecture into parallel and shared components, and it determines which GPU component can be effectively parallelized and can work correctly in multithreaded simulation. In addition, the work-group parallel simulation technique effectively boosts the performance of parallelized GPU simulation by eliminating the synchronization overhead. Experimental results obtained using a simulator with the proposed framework show that the proposed parallel simulation technique has a speed-up of up to 4.15 as compared to an existing sequential GPU simulator on an 8-core machine providing minimized cycle errors.

AB - GPU computing is at the forefront of high-performance computing, and it has greatly affected current studies on parallel software and hardware design because of its massively parallel architecture. Therefore, numerous studies have focused on the utilization of GPUs in various fields. However, studies of GPU architectures are constrained by the lack of a suitable GPU simulator. Previously proposed GPU simulators do not have sufficient simulation speed for advanced software and architecture studies. In this paper, we propose a new parallel simulation framework and a parallel simulation technique called work-group parallel simulation in order to improve the simulation speed for modern many-core GPUs. The proposed framework divides the GPU architecture into parallel and shared components, and it determines which GPU component can be effectively parallelized and can work correctly in multithreaded simulation. In addition, the work-group parallel simulation technique effectively boosts the performance of parallelized GPU simulation by eliminating the synchronization overhead. Experimental results obtained using a simulator with the proposed framework show that the proposed parallel simulation technique has a speed-up of up to 4.15 as compared to an existing sequential GPU simulator on an 8-core machine providing minimized cycle errors.

UR - http://www.scopus.com/inward/record.url?scp=84881446418&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84881446418&partnerID=8YFLogxK

U2 - 10.1109/ISPASS.2013.6557151

DO - 10.1109/ISPASS.2013.6557151

M3 - Conference contribution

SN - 9781467357777

SP - 107

EP - 117

BT - ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software

ER -

Lee S, Ro WW. Parallel GPU architecture simulation framework exploiting work allocation unit parallelism. In ISPASS 2013 - IEEE International Symposium on Performance Analysis of Systems and Software. 2013. p. 107-117. 6557151 https://doi.org/10.1109/ISPASS.2013.6557151