A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems

Volker Hampel; Grigori Goronzy; Erik Maehle

doi:10.1007/978-3-642-19137-4_1

A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems

Volker Hampel, Grigori Goronzy, Erik Maehle

Institute of Computer Engineering

Institute of Computer Engineering

Abstract

Special hardware accelerators like FPGAs and GPUs are commonly introduced into a computing system as a separate device. Consequently, the accelerator and the host system do not share a common memory. Sourcing out the data to the additional hardware thus introduces a communication penalty. Based on a combination of a program's source code and execution profiling we perform an analysis which evaluates the arithmetic intensity as a cost function to identify those parts most reasonable to source out to the accelerating hardware. The basic principles of this analysis are introduced and tested with a sample application. Its concrete results are discussed and evaluated based on the performance of a FPGA-based and a GPU-based implementation.

Original language	English
Title of host publication	ARCS 2011: Architecture of Computing Systems - ARCS 2011
Number of pages	12
Volume	6566
Publisher	Springer Verlag
Publication date	02.03.2011
Pages	1-12
ISBN (Print)	978-3-642-19136-7
ISBN (Electronic)	978-3-642-19137-4
DOIs	https://doi.org/10.1007/978-3-642-19137-4_1
Publication status	Published - 02.03.2011
Event	24th International Conference on Architecture of Computing Systems - Como, Italy Duration: 24.02.2011 → 25.02.2011 Conference number: 83943

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/978-3-642-19137-4_1

Cite this

@inproceedings{1278e6a7fe0849ebbe1aafdca13cb956,

title = "A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems",

abstract = "Special hardware accelerators like FPGAs and GPUs are commonly introduced into a computing system as a separate device. Consequently, the accelerator and the host system do not share a common memory. Sourcing out the data to the additional hardware thus introduces a communication penalty. Based on a combination of a program's source code and execution profiling we perform an analysis which evaluates the arithmetic intensity as a cost function to identify those parts most reasonable to source out to the accelerating hardware. The basic principles of this analysis are introduced and tested with a sample application. Its concrete results are discussed and evaluated based on the performance of a FPGA-based and a GPU-based implementation.",

author = "Volker Hampel and Grigori Goronzy and Erik Maehle",

year = "2011",

month = mar,

day = "2",

doi = "10.1007/978-3-642-19137-4_1",

language = "English",

isbn = "978-3-642-19136-7",

volume = "6566 ",

series = "Lecture Notes in Computer Science",

publisher = "Springer Verlag",

pages = "1--12",

booktitle = "ARCS 2011: Architecture of Computing Systems - ARCS 2011",

note = "24th International Conference on Architecture of Computing Systems <br/>, ARCS 2011 ; Conference date: 24-02-2011 Through 25-02-2011",

}

Hampel, V, Goronzy, G & Maehle, E 2011, A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems. in ARCS 2011: Architecture of Computing Systems - ARCS 2011 . vol. 6566 , Lecture Notes in Computer Science, vol. 6566 LNCS, Springer Verlag, pp. 1-12, 24th International Conference on Architecture of Computing Systems
, Como, Italy, 24.02.11. https://doi.org/10.1007/978-3-642-19137-4_1

A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems. / Hampel, Volker; Goronzy, Grigori; Maehle, Erik.
ARCS 2011: Architecture of Computing Systems - ARCS 2011 . Vol. 6566 Springer Verlag, 2011. p. 1-12 (Lecture Notes in Computer Science; Vol. 6566 LNCS).

Research output: Chapters in Books/Reports/Conference Proceedings › Conference contribution › peer-review

TY - GEN

T1 - A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems

AU - Hampel, Volker

AU - Goronzy, Grigori

AU - Maehle, Erik

N1 - Conference code: 83943

PY - 2011/3/2

Y1 - 2011/3/2

N2 - Special hardware accelerators like FPGAs and GPUs are commonly introduced into a computing system as a separate device. Consequently, the accelerator and the host system do not share a common memory. Sourcing out the data to the additional hardware thus introduces a communication penalty. Based on a combination of a program's source code and execution profiling we perform an analysis which evaluates the arithmetic intensity as a cost function to identify those parts most reasonable to source out to the accelerating hardware. The basic principles of this analysis are introduced and tested with a sample application. Its concrete results are discussed and evaluated based on the performance of a FPGA-based and a GPU-based implementation.

AB - Special hardware accelerators like FPGAs and GPUs are commonly introduced into a computing system as a separate device. Consequently, the accelerator and the host system do not share a common memory. Sourcing out the data to the additional hardware thus introduces a communication penalty. Based on a combination of a program's source code and execution profiling we perform an analysis which evaluates the arithmetic intensity as a cost function to identify those parts most reasonable to source out to the accelerating hardware. The basic principles of this analysis are introduced and tested with a sample application. Its concrete results are discussed and evaluated based on the performance of a FPGA-based and a GPU-based implementation.

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-79952031210&origin=inward&txGid=b2051602b4c6bc54396e393dcdffa98e

U2 - 10.1007/978-3-642-19137-4_1

DO - 10.1007/978-3-642-19137-4_1

M3 - Conference contribution

SN - 978-3-642-19136-7

VL - 6566

T3 - Lecture Notes in Computer Science

SP - 1

EP - 12

BT - ARCS 2011: Architecture of Computing Systems - ARCS 2011

PB - Springer Verlag

T2 - 24th International Conference on Architecture of Computing Systems <br/>

Y2 - 24 February 2011 through 25 February 2011

ER -

A Code-Based Analytical Approach for Using Separate Device Coprocessors in Computing Systems

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this