Not in name alone: A memristive memory processing unit for real in-memory processing

Ameer Haj-Ali; Rotem Ben-Hur; Nimrod Wald; Ronny Ronen; Shahar Kvatinsky

doi:10.1109/MM.2018.053631137

Not in name alone: A memristive memory processing unit for real in-memory processing

Ameer Haj-Ali, Rotem Ben-Hur, Nimrod Wald, Ronny Ronen, Shahar Kvatinsky

Research output: Contribution to journal › Article › peer-review

24 Scopus citations

Abstract

Data movement between processing and memory is the root cause of the limited performance and energy efficiency in modern von Neumann systems. To overcome the data-movement bottleneck, we present the memristive Memory Processing Unit (mMPU)-a real processing-in-memory system in which the computation is done directly in the memory cells, thus eliminating the necessity for data transfer. Furthermore, with its enormous inner parallelism, this system is ideal for data-intensive applications that are based on single instruction, multiple data (SIMD)-providing high throughput and energy-efficiency.

Original language	English
Article number	8474943
Pages (from-to)	13-21
Number of pages	9
Journal	IEEE Micro
Volume	38
Issue number	5
DOIs	https://doi.org/10.1109/MM.2018.053631137
State	Published - 1 Sep 2018
Externally published	Yes

Keywords

MAGIC
hardware
memristor
von Neumann bottleneck

ASJC Scopus subject areas

Software
Hardware and Architecture
Electrical and Electronic Engineering

UN SDGs

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

Access to Document

10.1109/MM.2018.053631137

Cite this

@article{2f16949122b44c29b480f8dd243a277b,

title = "Not in name alone: A memristive memory processing unit for real in-memory processing",

abstract = "Data movement between processing and memory is the root cause of the limited performance and energy efficiency in modern von Neumann systems. To overcome the data-movement bottleneck, we present the memristive Memory Processing Unit (mMPU)-a real processing-in-memory system in which the computation is done directly in the memory cells, thus eliminating the necessity for data transfer. Furthermore, with its enormous inner parallelism, this system is ideal for data-intensive applications that are based on single instruction, multiple data (SIMD)-providing high throughput and energy-efficiency.",

keywords = "MAGIC, hardware, memristor, von Neumann bottleneck",

author = "Ameer Haj-Ali and Rotem Ben-Hur and Nimrod Wald and Ronny Ronen and Shahar Kvatinsky",

note = "Funding Information: This research is partially supported by the European Research Council under the European Union{\textquoteright}s Horizon 2020 Research and Innovation Programme (grant agreement no. 757259), by the Viterbi Fellowship at the Technion Computer Engineering Center, by the EU ICT COST Action IC1401, and by the Israel Science Foundation grant no. 1514/17. Publisher Copyright: {\textcopyright} 1981-2012 IEEE.",

year = "2018",

month = sep,

day = "1",

doi = "10.1109/MM.2018.053631137",

language = "English",

volume = "38",

pages = "13--21",

journal = "IEEE Micro",

issn = "0272-1732",

publisher = "Institute of Electrical and Electronics Engineers",

number = "5",

}

TY - JOUR

T1 - Not in name alone

T2 - A memristive memory processing unit for real in-memory processing

AU - Haj-Ali, Ameer

AU - Ben-Hur, Rotem

AU - Wald, Nimrod

AU - Ronen, Ronny

AU - Kvatinsky, Shahar

N1 - Funding Information: This research is partially supported by the European Research Council under the European Union’s Horizon 2020 Research and Innovation Programme (grant agreement no. 757259), by the Viterbi Fellowship at the Technion Computer Engineering Center, by the EU ICT COST Action IC1401, and by the Israel Science Foundation grant no. 1514/17. Publisher Copyright: © 1981-2012 IEEE.

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Data movement between processing and memory is the root cause of the limited performance and energy efficiency in modern von Neumann systems. To overcome the data-movement bottleneck, we present the memristive Memory Processing Unit (mMPU)-a real processing-in-memory system in which the computation is done directly in the memory cells, thus eliminating the necessity for data transfer. Furthermore, with its enormous inner parallelism, this system is ideal for data-intensive applications that are based on single instruction, multiple data (SIMD)-providing high throughput and energy-efficiency.

AB - Data movement between processing and memory is the root cause of the limited performance and energy efficiency in modern von Neumann systems. To overcome the data-movement bottleneck, we present the memristive Memory Processing Unit (mMPU)-a real processing-in-memory system in which the computation is done directly in the memory cells, thus eliminating the necessity for data transfer. Furthermore, with its enormous inner parallelism, this system is ideal for data-intensive applications that are based on single instruction, multiple data (SIMD)-providing high throughput and energy-efficiency.

KW - MAGIC

KW - hardware

KW - memristor

KW - von Neumann bottleneck

UR - http://www.scopus.com/inward/record.url?scp=85054530538&partnerID=8YFLogxK

U2 - 10.1109/MM.2018.053631137

DO - 10.1109/MM.2018.053631137

M3 - Article

AN - SCOPUS:85054530538

SN - 0272-1732

VL - 38

SP - 13

EP - 21

JO - IEEE Micro

JF - IEEE Micro

IS - 5

M1 - 8474943

ER -

Not in name alone: A memristive memory processing unit for real in-memory processing

Abstract

Keywords

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this