Simplicial Piecewise Linear Computation Complexity for Vector-Vector Products

P. Julian; M. Villemur

doi:10.1109/CAE48787.2020.9046379

Simplicial Piecewise Linear Computation Complexity for Vector-Vector Products

P. Julian, M. Villemur

Research output: Conference proceeding/Chapter in Book/Report/ › Conference Paper › peer-review

Abstract

In this paper we analyze several simplicial piecewise-linear digital computation architectures that are good candidates to replace the multiply and accumulate (MAC) units in deep neural networks. We analyze them in terms of area and energy of a digital VLSI realization and compare the results with the classic MAC architectures.

Original language	English
Title of host publication	2020 Argentine Conference on Electronics (CAE)
Pages	108-113
Number of pages	6
DOIs	https://doi.org/10.1109/CAE48787.2020.9046379
Publication status	Published - 28 Feb 2020
Externally published	Yes
Event	2020 Argentine Conference on Electronics (CAE) - Buenos Aires, Argentina Duration: 27 Feb 2020 → 28 Feb 2020

Publication series

Name	2020 Argentine Conference on Electronics (CAE)

Conference

Conference	2020 Argentine Conference on Electronics (CAE)
Period	27/02/20 → 28/02/20

Keywords

Linear combination
piecewise-linear functions
vector-vector multiplication

Access to Document

10.1109/CAE48787.2020.9046379

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9046379

Cite this

@inproceedings{286aeb44d90647c9983db98c8b84b557,

title = "Simplicial Piecewise Linear Computation Complexity for Vector-Vector Products",

abstract = "In this paper we analyze several simplicial piecewise-linear digital computation architectures that are good candidates to replace the multiply and accumulate (MAC) units in deep neural networks. We analyze them in terms of area and energy of a digital VLSI realization and compare the results with the classic MAC architectures.",

keywords = "Linear combination, piecewise-linear functions, vector-vector multiplication",

author = "P. Julian and M. Villemur",

year = "2020",

month = feb,

day = "28",

doi = "10.1109/CAE48787.2020.9046379",

language = "English",

isbn = "978-1-7281-6196-9",

series = "2020 Argentine Conference on Electronics (CAE)",

pages = "108--113",

booktitle = "2020 Argentine Conference on Electronics (CAE)",

note = "2020 Argentine Conference on Electronics (CAE) ; Conference date: 27-02-2020 Through 28-02-2020",

}

TY - GEN

T1 - Simplicial Piecewise Linear Computation Complexity for Vector-Vector Products

AU - Julian, P.

AU - Villemur, M.

PY - 2020/2/28

Y1 - 2020/2/28

N2 - In this paper we analyze several simplicial piecewise-linear digital computation architectures that are good candidates to replace the multiply and accumulate (MAC) units in deep neural networks. We analyze them in terms of area and energy of a digital VLSI realization and compare the results with the classic MAC architectures.

AB - In this paper we analyze several simplicial piecewise-linear digital computation architectures that are good candidates to replace the multiply and accumulate (MAC) units in deep neural networks. We analyze them in terms of area and energy of a digital VLSI realization and compare the results with the classic MAC architectures.

KW - Linear combination

KW - piecewise-linear functions

KW - vector-vector multiplication

UR - https://ieeexplore.ieee.org/document/9046379/

U2 - 10.1109/CAE48787.2020.9046379

DO - 10.1109/CAE48787.2020.9046379

M3 - Conference Paper

SN - 978-1-7281-6196-9

T3 - 2020 Argentine Conference on Electronics (CAE)

SP - 108

EP - 113

BT - 2020 Argentine Conference on Electronics (CAE)

T2 - 2020 Argentine Conference on Electronics (CAE)

Y2 - 27 February 2020 through 28 February 2020

ER -

Simplicial Piecewise Linear Computation Complexity for Vector-Vector Products

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this