TDO-CIM: Transparent Detection and Offloading for Computation In-memory

Kanishkan Vadivel; Lorenzo Chelini; Ali BanaGozar; Gagandeep Singh; Stefano Corda; Roel Jordans; Henk Corporaal

doi:10.23919/DATE48585.2020.9116464

TDO-CIM: Transparent Detection and Offloading for Computation In-memory

Kanishkan Vadivel
, Lorenzo Chelini
, Ali BanaGozar
, Gagandeep Singh
, Stefano Corda
, Roel Jordans
, Henk Corporaal

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

15 Citations (Scopus)

126 Downloads (Pure)

Abstract

Computation in-memory is a promising non-von Neumann approach aiming at completely diminishing the data transfer to and from the memory subsystem. Although a lot of architectures have been proposed, compiler support for such architectures is still lagging behind. In this paper, we close this gap by proposing an end-to-end compilation flow for in-memory computing based on the LLVM compiler infrastructure. Starting from sequential code, our approach automatically detects, optimizes, and offloads kernels suitable for in-memory acceleration. We demonstrate our compiler tool-flow on the PolyBench/C benchmark suite and evaluate the benefits of our proposed in-memory architecture simulated in Gem5 by comparing it with a state-of-the-art von Neumann architecture.

Original language	English
Title of host publication	2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)
Editors	Giorgio Di Natale, Cristiana Bolchini, Elena-Ioana Vatajelu
Publisher	Institute of Electrical and Electronics Engineers
Pages	1602-1605
Number of pages	4
ISBN (Electronic)	978-3-9819263-4-7
DOIs	https://doi.org/10.23919/DATE48585.2020.9116464
Publication status	Published - 30 Jun 2020
Event	23rd Design, Automation and Test in Europe Conference and Exhibition (DATE 2020) - Grenoble, France Duration: 9 Mar 2020 → 13 Mar 2020 Conference number: 23

Conference

Conference	23rd Design, Automation and Test in Europe Conference and Exhibition (DATE 2020)
Abbreviated title	DATE 2020
Country/Territory	France
City	Grenoble
Period	9/03/20 → 13/03/20

Funding

Funders	Funder number
European Union's Horizon 2020 - Research and Innovation Framework Programme
European Union's Horizon 2020 - Research and Innovation Framework Programme	780215, 676240

Keywords

compute in memory
LLVM
Loop Tactics
memristor
pattern matching
Polly

Access to Document

10.23919/DATE48585.2020.9116464

arXiv 2007.00060v1Final author version , 520 KB

Cite this

Vadivel, K., Chelini, L., BanaGozar, A., Singh, G., Corda, S., Jordans, R., & Corporaal, H. (2020). TDO-CIM: Transparent Detection and Offloading for Computation In-memory. In G. Di Natale, C. Bolchini, & E.-I. Vatajelu (Eds.), 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE) (pp. 1602-1605). Article 9116464 Institute of Electrical and Electronics Engineers. https://doi.org/10.23919/DATE48585.2020.9116464

@inproceedings{c17d6ddbcfc94039950c3c066d0b6645,

title = "TDO-CIM: Transparent Detection and Offloading for Computation In-memory",

abstract = " Computation in-memory is a promising non-von Neumann approach aiming at completely diminishing the data transfer to and from the memory subsystem. Although a lot of architectures have been proposed, compiler support for such architectures is still lagging behind. In this paper, we close this gap by proposing an end-to-end compilation flow for in-memory computing based on the LLVM compiler infrastructure. Starting from sequential code, our approach automatically detects, optimizes, and offloads kernels suitable for in-memory acceleration. We demonstrate our compiler tool-flow on the PolyBench/C benchmark suite and evaluate the benefits of our proposed in-memory architecture simulated in Gem5 by comparing it with a state-of-the-art von Neumann architecture. ",

keywords = "cs.AR, cs.ET, compute in memory, LLVM, Loop Tactics, memristor, pattern matching, Polly",

author = "Kanishkan Vadivel and Lorenzo Chelini and Ali BanaGozar and Gagandeep Singh and Stefano Corda and Roel Jordans and Henk Corporaal",

year = "2020",

month = jun,

day = "30",

doi = "10.23919/DATE48585.2020.9116464",

language = "English",

pages = "1602--1605",

editor = "\{Di Natale\}, Giorgio and Cristiana Bolchini and Elena-Ioana Vatajelu",

booktitle = "2020 Design, Automation \& Test in Europe Conference \& Exhibition (DATE)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "23rd Design, Automation and Test in Europe Conference and Exhibition (DATE 2020), DATE 2020 ; Conference date: 09-03-2020 Through 13-03-2020",

}

Vadivel, K, Chelini, L, BanaGozar, A, Singh, G, Corda, S, Jordans, R & Corporaal, H 2020, TDO-CIM: Transparent Detection and Offloading for Computation In-memory. in G Di Natale, C Bolchini & E-I Vatajelu (eds), 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)., 9116464, Institute of Electrical and Electronics Engineers, pp. 1602-1605, 23rd Design, Automation and Test in Europe Conference and Exhibition (DATE 2020), Grenoble, France, 9/03/20. https://doi.org/10.23919/DATE48585.2020.9116464

TDO-CIM: Transparent Detection and Offloading for Computation In-memory. / Vadivel, Kanishkan; Chelini, Lorenzo; BanaGozar, Ali et al.
2020 Design, Automation & Test in Europe Conference & Exhibition (DATE). ed. / Giorgio Di Natale; Cristiana Bolchini; Elena-Ioana Vatajelu. Institute of Electrical and Electronics Engineers, 2020. p. 1602-1605 9116464.

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

TY - GEN

T1 - TDO-CIM

T2 - 23rd Design, Automation and Test in Europe Conference and Exhibition (DATE 2020)

AU - Vadivel, Kanishkan

AU - Chelini, Lorenzo

AU - BanaGozar, Ali

AU - Singh, Gagandeep

AU - Corda, Stefano

AU - Jordans, Roel

AU - Corporaal, Henk

N1 - Conference code: 23

PY - 2020/6/30

Y1 - 2020/6/30

N2 - Computation in-memory is a promising non-von Neumann approach aiming at completely diminishing the data transfer to and from the memory subsystem. Although a lot of architectures have been proposed, compiler support for such architectures is still lagging behind. In this paper, we close this gap by proposing an end-to-end compilation flow for in-memory computing based on the LLVM compiler infrastructure. Starting from sequential code, our approach automatically detects, optimizes, and offloads kernels suitable for in-memory acceleration. We demonstrate our compiler tool-flow on the PolyBench/C benchmark suite and evaluate the benefits of our proposed in-memory architecture simulated in Gem5 by comparing it with a state-of-the-art von Neumann architecture.

AB - Computation in-memory is a promising non-von Neumann approach aiming at completely diminishing the data transfer to and from the memory subsystem. Although a lot of architectures have been proposed, compiler support for such architectures is still lagging behind. In this paper, we close this gap by proposing an end-to-end compilation flow for in-memory computing based on the LLVM compiler infrastructure. Starting from sequential code, our approach automatically detects, optimizes, and offloads kernels suitable for in-memory acceleration. We demonstrate our compiler tool-flow on the PolyBench/C benchmark suite and evaluate the benefits of our proposed in-memory architecture simulated in Gem5 by comparing it with a state-of-the-art von Neumann architecture.

KW - cs.AR

KW - cs.ET

KW - compute in memory

KW - LLVM

KW - Loop Tactics

KW - memristor

KW - pattern matching

KW - Polly

UR - https://www.scopus.com/pages/publications/85087430237

U2 - 10.23919/DATE48585.2020.9116464

DO - 10.23919/DATE48585.2020.9116464

M3 - Conference contribution

SP - 1602

EP - 1605

BT - 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)

A2 - Di Natale, Giorgio

A2 - Bolchini, Cristiana

A2 - Vatajelu, Elena-Ioana

PB - Institute of Electrical and Electronics Engineers

Y2 - 9 March 2020 through 13 March 2020

ER -

Vadivel K, Chelini L, BanaGozar A, Singh G, Corda S, Jordans R et al. TDO-CIM: Transparent Detection and Offloading for Computation In-memory. In Di Natale G, Bolchini C, Vatajelu EI, editors, 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE). Institute of Electrical and Electronics Engineers. 2020. p. 1602-1605. 9116464 doi: 10.23919/DATE48585.2020.9116464

TDO-CIM: Transparent Detection and Offloading for Computation In-memory

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

MNEMOSENE - Computation-in-memory architecture based on resistive devices

Cite this

TDO-CIM: Transparent Detection and Offloading for Computation In-memory

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Projects

MNEMOSENE - Computation-in-memory architecture based on resistive devices

Cite this