System Simulation of Memristor Based Computation in Memory Platforms

Ali BanaGozar; Kanishkan Vadivel; Joonas Multanen; Pekka Jääskeläinen; Sander Stuijk; Henk Corporaal

doi:10.1007/978-3-030-60939-9_11

System Simulation of Memristor Based Computation in Memory Platforms

Ali BanaGozar, Kanishkan Vadivel, Joonas Multanen, Pekka Jääskeläinen, Sander Stuijk, Henk Corporaal

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

8 Citations (Scopus)

Abstract

Processors based on the von Neumann architecture show inefficient performance on many emerging data-intensive workloads. Computation in-memory (CIM) tries to address this challenge by performing the computation on the data location. To realize CIM, memristors, that are deployed in a crossbar structure, are a promising candidate. Even though extensive research has been carried out on memristors at device/circuit-level, the implications of their integration as accelerators (CIM units) in a full-blown system are not studied extensively. To study that, we developed a simulator for memristor crossbar and its analog peripheries. This paper evaluates a complete system consisting of a Transport Triggered Architecture (TTA) based host core integrating one or more CIM units. This evaluation is based on a cycle-accurate simulation. For this purpose we designed a simulator which a) includes the memristor crossbar operations as well as its surrounding analog drivers, b) provides the required interface to the co-processing digital elements, and c) presents a micro-instruction set architecture (micro-ISA) that controls and operates both analog and digital components. It is used to assess the effectiveness of the CIM unit in terms of performance, energy, and area in a full-blown system. It is shown, for example, that the EDAP for the deep learning application, LeNet, is reduced by 84% in a full-blown system deploying memristor based crossbars.

Original language	English
Title of host publication	Embedded Computer Systems
Subtitle of host publication	Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings
Editors	Alex Orailoglu, Matthias Jung, Marc Reichenbach
Publisher	Springer
Pages	152-168
Number of pages	17
ISBN (Electronic)	978-3-030-60939-9
ISBN (Print)	978-3-030-60938-2
DOIs	https://doi.org/10.1007/978-3-030-60939-9_11
Publication status	Published - 2020
Event	20th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, SAMOS 2020 - Samos, Greece Duration: 5 Jul 2020 → 9 Jul 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12471 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	20th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, SAMOS 2020
Country/Territory	Greece
City	Samos
Period	5/07/20 → 9/07/20

Funding

Acknowledgment. This research is supported by EC Horizon 2020 Research and Innovation Program through MNEMOSENE project under Grant 780215. The work also recevied support from the FitOptiVis project [1] funded by the ECSEL Joint Undertaking under grant number H2020-ECSEL-2017-2-783162.

Funders	Funder number
European Union's Horizon 2020 - Research and Innovation Framework Programme	H2020-ECSEL-2017-2-783162
European Union's Horizon 2020 - Research and Innovation Framework Programme	780215

Keywords

Computation in memory
Memristor
Non-von neumann architectures
Simulator
Transport Triggered Architecture

Access to Document

10.1007/978-3-030-60939-9_11

Cite this

BanaGozar, A., Vadivel, K., Multanen, J., Jääskeläinen, P., Stuijk, S., & Corporaal, H. (2020). System Simulation of Memristor Based Computation in Memory Platforms. In A. Orailoglu, M. Jung, & M. Reichenbach (Eds.), Embedded Computer Systems: Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings (pp. 152-168). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12471 LNCS). Springer. https://doi.org/10.1007/978-3-030-60939-9_11

BanaGozar, Ali ; Vadivel, Kanishkan ; Multanen, Joonas et al. / System Simulation of Memristor Based Computation in Memory Platforms. Embedded Computer Systems: Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings. editor / Alex Orailoglu ; Matthias Jung ; Marc Reichenbach. Springer, 2020. pp. 152-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{ea6cc1cfeca04655a923375010447e8a,

title = "System Simulation of Memristor Based Computation in Memory Platforms",

abstract = "Processors based on the von Neumann architecture show inefficient performance on many emerging data-intensive workloads. Computation in-memory (CIM) tries to address this challenge by performing the computation on the data location. To realize CIM, memristors, that are deployed in a crossbar structure, are a promising candidate. Even though extensive research has been carried out on memristors at device/circuit-level, the implications of their integration as accelerators (CIM units) in a full-blown system are not studied extensively. To study that, we developed a simulator for memristor crossbar and its analog peripheries. This paper evaluates a complete system consisting of a Transport Triggered Architecture (TTA) based host core integrating one or more CIM units. This evaluation is based on a cycle-accurate simulation. For this purpose we designed a simulator which a) includes the memristor crossbar operations as well as its surrounding analog drivers, b) provides the required interface to the co-processing digital elements, and c) presents a micro-instruction set architecture (micro-ISA) that controls and operates both analog and digital components. It is used to assess the effectiveness of the CIM unit in terms of performance, energy, and area in a full-blown system. It is shown, for example, that the EDAP for the deep learning application, LeNet, is reduced by 84% in a full-blown system deploying memristor based crossbars.",

keywords = "Computation in memory, Memristor, Non-von neumann architectures, Simulator, Transport Triggered Architecture",

author = "Ali BanaGozar and Kanishkan Vadivel and Joonas Multanen and Pekka J{\"a}{\"a}skel{\"a}inen and Sander Stuijk and Henk Corporaal",

year = "2020",

doi = "10.1007/978-3-030-60939-9_11",

language = "English",

isbn = "978-3-030-60938-2",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "152--168",

editor = "Alex Orailoglu and Matthias Jung and Marc Reichenbach",

booktitle = "Embedded Computer Systems",

address = "Germany",

note = "20th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, SAMOS 2020 ; Conference date: 05-07-2020 Through 09-07-2020",

}

BanaGozar, A , Vadivel, K, Multanen, J, Jääskeläinen, P, Stuijk, S & Corporaal, H 2020, System Simulation of Memristor Based Computation in Memory Platforms. in A Orailoglu, M Jung & M Reichenbach (eds), Embedded Computer Systems: Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12471 LNCS, Springer, pp. 152-168, 20th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, SAMOS 2020, Samos, Greece, 5/07/20. https://doi.org/10.1007/978-3-030-60939-9_11

System Simulation of Memristor Based Computation in Memory Platforms. / BanaGozar, Ali ; Vadivel, Kanishkan; Multanen, Joonas et al.
Embedded Computer Systems: Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings. ed. / Alex Orailoglu; Matthias Jung; Marc Reichenbach. Springer, 2020. p. 152-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12471 LNCS).

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

TY - GEN

T1 - System Simulation of Memristor Based Computation in Memory Platforms

AU - BanaGozar, Ali

AU - Vadivel, Kanishkan

AU - Multanen, Joonas

AU - Jääskeläinen, Pekka

AU - Stuijk, Sander

AU - Corporaal, Henk

PY - 2020

Y1 - 2020

N2 - Processors based on the von Neumann architecture show inefficient performance on many emerging data-intensive workloads. Computation in-memory (CIM) tries to address this challenge by performing the computation on the data location. To realize CIM, memristors, that are deployed in a crossbar structure, are a promising candidate. Even though extensive research has been carried out on memristors at device/circuit-level, the implications of their integration as accelerators (CIM units) in a full-blown system are not studied extensively. To study that, we developed a simulator for memristor crossbar and its analog peripheries. This paper evaluates a complete system consisting of a Transport Triggered Architecture (TTA) based host core integrating one or more CIM units. This evaluation is based on a cycle-accurate simulation. For this purpose we designed a simulator which a) includes the memristor crossbar operations as well as its surrounding analog drivers, b) provides the required interface to the co-processing digital elements, and c) presents a micro-instruction set architecture (micro-ISA) that controls and operates both analog and digital components. It is used to assess the effectiveness of the CIM unit in terms of performance, energy, and area in a full-blown system. It is shown, for example, that the EDAP for the deep learning application, LeNet, is reduced by 84% in a full-blown system deploying memristor based crossbars.

AB - Processors based on the von Neumann architecture show inefficient performance on many emerging data-intensive workloads. Computation in-memory (CIM) tries to address this challenge by performing the computation on the data location. To realize CIM, memristors, that are deployed in a crossbar structure, are a promising candidate. Even though extensive research has been carried out on memristors at device/circuit-level, the implications of their integration as accelerators (CIM units) in a full-blown system are not studied extensively. To study that, we developed a simulator for memristor crossbar and its analog peripheries. This paper evaluates a complete system consisting of a Transport Triggered Architecture (TTA) based host core integrating one or more CIM units. This evaluation is based on a cycle-accurate simulation. For this purpose we designed a simulator which a) includes the memristor crossbar operations as well as its surrounding analog drivers, b) provides the required interface to the co-processing digital elements, and c) presents a micro-instruction set architecture (micro-ISA) that controls and operates both analog and digital components. It is used to assess the effectiveness of the CIM unit in terms of performance, energy, and area in a full-blown system. It is shown, for example, that the EDAP for the deep learning application, LeNet, is reduced by 84% in a full-blown system deploying memristor based crossbars.

KW - Computation in memory

KW - Memristor

KW - Non-von neumann architectures

KW - Simulator

KW - Transport Triggered Architecture

UR - http://www.scopus.com/inward/record.url?scp=85093857229&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-60939-9_11

DO - 10.1007/978-3-030-60939-9_11

M3 - Conference contribution

AN - SCOPUS:85093857229

SN - 978-3-030-60938-2

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 152

EP - 168

BT - Embedded Computer Systems

A2 - Orailoglu, Alex

A2 - Jung, Matthias

A2 - Reichenbach, Marc

PB - Springer

T2 - 20th International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, SAMOS 2020

Y2 - 5 July 2020 through 9 July 2020

ER -

BanaGozar A , Vadivel K, Multanen J, Jääskeläinen P, Stuijk S , Corporaal H. System Simulation of Memristor Based Computation in Memory Platforms. In Orailoglu A, Jung M, Reichenbach M, editors, Embedded Computer Systems: Architectures, Modeling, and Simulation - 20th International Conference, SAMOS 2020, Proceedings. Springer. 2020. p. 152-168. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-60939-9_11

System Simulation of Memristor Based Computation in Memory Platforms

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

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

Cite this

System Simulation of Memristor Based Computation in Memory Platforms

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Projects

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

Cite this