A study of recent contributions on performance and simulation techniques for accelerator devices

Hussein Ajam; Michael Opoku Agyeman

doi:10.1109/ICEEE2.2017.7935829

A study of recent contributions on performance and simulation techniques for accelerator devices

Research output: Contribution to Book/Report › Conference Contribution › peer-review

Abstract

High performance computing platform is moving from homogeneous individual unites to heterogeneous systems where each unit is a combination of homogeneous cores and accelerator devices. Accelerators such as GPUs, FPGAs, DSPs, are usually designed for specific and intensive type of computing tasks. The presence of these devises have created fresh and attractive development platforms for developers and designers as well as novel performance analysis frameworks and optimization tools. This is the cutting edge in performance of some accelerator devices like: GPUs and Intel’s Xeon Phi. We outline some of the existing heterogeneous systems and their development frameworks. The core of this study is a review of performance modeling of these devices.

Original language	English
Title of host publication	International Conference on Electrical and Electronics Engineering
Place of Publication	Turkey
Publisher	IEEE
Pages	256-260
Number of pages	5
ISBN (Electronic)	978-1-5090-6789-3
ISBN (Print)	978-1-5090-6790-9
DOIs	https://doi.org/10.1109/ICEEE2.2017.7935829
Publication status	Published - 1 Jun 2017
Event	2017 4th International Conference on Electrical and Electronics Engineering - Ankara, Turkey Duration: 8 Apr 2017 → 10 Apr 2017

Conference

Conference	2017 4th International Conference on Electrical and Electronics Engineering
Country/Territory	Turkey
City	Ankara
Period	8/04/17 → 10/04/17

Keywords

GPU
FPGA
DSP
GPGPU

Access to Document

10.1109/ICEEE2.2017.7935829

Cite this

@inproceedings{6844e5cd2a6049c9a850359829ef2f46,

title = "A study of recent contributions on performance and simulation techniques for accelerator devices",

abstract = "High performance computing platform is moving from homogeneous individual unites to heterogeneous systems where each unit is a combination of homogeneous cores and accelerator devices. Accelerators such as GPUs, FPGAs, DSPs, are usually designed for specific and intensive type of computing tasks. The presence of these devises have created fresh and attractive development platforms for developers and designers as well as novel performance analysis frameworks and optimization tools. This is the cutting edge in performance of some accelerator devices like: GPUs and Intel{\textquoteright}s Xeon Phi. We outline some of the existing heterogeneous systems and their development frameworks. The core of this study is a review of performance modeling of these devices.",

keywords = "GPU, FPGA, DSP, GPGPU",

author = "Hussein Ajam and {Opoku Agyeman}, Michael",

year = "2017",

month = jun,

day = "1",

doi = "10.1109/ICEEE2.2017.7935829",

language = "English",

isbn = "978-1-5090-6790-9",

pages = "256--260",

booktitle = "International Conference on Electrical and Electronics Engineering",

publisher = "IEEE",

note = "2017 4th International Conference on Electrical and Electronics Engineering ; Conference date: 08-04-2017 Through 10-04-2017",

}

Ajam, H & Opoku Agyeman, M 2017, A study of recent contributions on performance and simulation techniques for accelerator devices. in International Conference on Electrical and Electronics Engineering. IEEE, Turkey, pp. 256-260, 2017 4th International Conference on Electrical and Electronics Engineering, Ankara, Turkey, 8/04/17. https://doi.org/10.1109/ICEEE2.2017.7935829

TY - GEN

T1 - A study of recent contributions on performance and simulation techniques for accelerator devices

AU - Ajam, Hussein

AU - Opoku Agyeman, Michael

PY - 2017/6/1

Y1 - 2017/6/1

N2 - High performance computing platform is moving from homogeneous individual unites to heterogeneous systems where each unit is a combination of homogeneous cores and accelerator devices. Accelerators such as GPUs, FPGAs, DSPs, are usually designed for specific and intensive type of computing tasks. The presence of these devises have created fresh and attractive development platforms for developers and designers as well as novel performance analysis frameworks and optimization tools. This is the cutting edge in performance of some accelerator devices like: GPUs and Intel’s Xeon Phi. We outline some of the existing heterogeneous systems and their development frameworks. The core of this study is a review of performance modeling of these devices.

AB - High performance computing platform is moving from homogeneous individual unites to heterogeneous systems where each unit is a combination of homogeneous cores and accelerator devices. Accelerators such as GPUs, FPGAs, DSPs, are usually designed for specific and intensive type of computing tasks. The presence of these devises have created fresh and attractive development platforms for developers and designers as well as novel performance analysis frameworks and optimization tools. This is the cutting edge in performance of some accelerator devices like: GPUs and Intel’s Xeon Phi. We outline some of the existing heterogeneous systems and their development frameworks. The core of this study is a review of performance modeling of these devices.

KW - GPU

KW - FPGA

KW - DSP

KW - GPGPU

U2 - 10.1109/ICEEE2.2017.7935829

DO - 10.1109/ICEEE2.2017.7935829

M3 - Conference Contribution

SN - 978-1-5090-6790-9

SP - 256

EP - 260

BT - International Conference on Electrical and Electronics Engineering

PB - IEEE

CY - Turkey

T2 - 2017 4th International Conference on Electrical and Electronics Engineering

Y2 - 8 April 2017 through 10 April 2017

ER -

A study of recent contributions on performance and simulation techniques for accelerator devices

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this