Symmetric 2-d-memory access to multidimensional data

Sumitha George, Xueqing Li, Minli Julie Liao, Kaisheng Ma, Srivatsa Srinivasa, Karthik Mohan, Ahmedullah Aziz, John Morgan Sampson, Sumeet Kumar Gupta, Vijaykrishnan Narayanan

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

In this paper, we propose a novel memory architecture with the capability of single-cycle row-wise/column-wise accesses. Such an architecture is highly suitable for workloads featuring spatial locality in multiple dimensions, which is a characteristic of many matrix and array operations. We describe in detail the circuit design techniques enabling the proposed architectures, as well as the viability of emerging memory technologies based on ferroelectric transistors (FEFETs) for our design. Compared to FEFET memory with standard 1-D access, we achieve 5% energy savings for the proposed memory featuring 2-D read and 93% energy savings for memory with 2-D read and write, for 32 bit column read and write. In addition, we get around 11% and 95% delay savings for 2-D read-enabled memory and 2-D read-write memory, respectively. The application analysis shows that 2-D read-enabled memory achieves around 86% average decrease in row-buffer transactions in 256 × 256 size matrix operations without any array area increase. The 2-D read write memory offers 87% decrease in row-buffer transactions with 28.5% increase in array area compared to the 1-D FEFET memory.

Original languageEnglish (US)
Pages (from-to)1040-1050
Number of pages11
JournalIEEE Transactions on Very Large Scale Integration (VLSI) Systems
Volume26
Issue number6
DOIs
StatePublished - Jun 1 2018

Fingerprint

Data storage equipment
Energy conservation
Memory architecture
Ferroelectric materials
Transistors
Networks (circuits)

All Science Journal Classification (ASJC) codes

  • Software
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

George, Sumitha ; Li, Xueqing ; Liao, Minli Julie ; Ma, Kaisheng ; Srinivasa, Srivatsa ; Mohan, Karthik ; Aziz, Ahmedullah ; Sampson, John Morgan ; Gupta, Sumeet Kumar ; Narayanan, Vijaykrishnan. / Symmetric 2-d-memory access to multidimensional data. In: IEEE Transactions on Very Large Scale Integration (VLSI) Systems. 2018 ; Vol. 26, No. 6. pp. 1040-1050.
@article{9d7aebabfdfd4b36bd221c32f9515028,
title = "Symmetric 2-d-memory access to multidimensional data",
abstract = "In this paper, we propose a novel memory architecture with the capability of single-cycle row-wise/column-wise accesses. Such an architecture is highly suitable for workloads featuring spatial locality in multiple dimensions, which is a characteristic of many matrix and array operations. We describe in detail the circuit design techniques enabling the proposed architectures, as well as the viability of emerging memory technologies based on ferroelectric transistors (FEFETs) for our design. Compared to FEFET memory with standard 1-D access, we achieve 5{\%} energy savings for the proposed memory featuring 2-D read and 93{\%} energy savings for memory with 2-D read and write, for 32 bit column read and write. In addition, we get around 11{\%} and 95{\%} delay savings for 2-D read-enabled memory and 2-D read-write memory, respectively. The application analysis shows that 2-D read-enabled memory achieves around 86{\%} average decrease in row-buffer transactions in 256 × 256 size matrix operations without any array area increase. The 2-D read write memory offers 87{\%} decrease in row-buffer transactions with 28.5{\%} increase in array area compared to the 1-D FEFET memory.",
author = "Sumitha George and Xueqing Li and Liao, {Minli Julie} and Kaisheng Ma and Srivatsa Srinivasa and Karthik Mohan and Ahmedullah Aziz and Sampson, {John Morgan} and Gupta, {Sumeet Kumar} and Vijaykrishnan Narayanan",
year = "2018",
month = "6",
day = "1",
doi = "10.1109/TVLSI.2018.2801302",
language = "English (US)",
volume = "26",
pages = "1040--1050",
journal = "IEEE Transactions on Very Large Scale Integration (VLSI) Systems",
issn = "1063-8210",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

Symmetric 2-d-memory access to multidimensional data. / George, Sumitha; Li, Xueqing; Liao, Minli Julie; Ma, Kaisheng; Srinivasa, Srivatsa; Mohan, Karthik; Aziz, Ahmedullah; Sampson, John Morgan; Gupta, Sumeet Kumar; Narayanan, Vijaykrishnan.

In: IEEE Transactions on Very Large Scale Integration (VLSI) Systems, Vol. 26, No. 6, 01.06.2018, p. 1040-1050.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Symmetric 2-d-memory access to multidimensional data

AU - George, Sumitha

AU - Li, Xueqing

AU - Liao, Minli Julie

AU - Ma, Kaisheng

AU - Srinivasa, Srivatsa

AU - Mohan, Karthik

AU - Aziz, Ahmedullah

AU - Sampson, John Morgan

AU - Gupta, Sumeet Kumar

AU - Narayanan, Vijaykrishnan

PY - 2018/6/1

Y1 - 2018/6/1

N2 - In this paper, we propose a novel memory architecture with the capability of single-cycle row-wise/column-wise accesses. Such an architecture is highly suitable for workloads featuring spatial locality in multiple dimensions, which is a characteristic of many matrix and array operations. We describe in detail the circuit design techniques enabling the proposed architectures, as well as the viability of emerging memory technologies based on ferroelectric transistors (FEFETs) for our design. Compared to FEFET memory with standard 1-D access, we achieve 5% energy savings for the proposed memory featuring 2-D read and 93% energy savings for memory with 2-D read and write, for 32 bit column read and write. In addition, we get around 11% and 95% delay savings for 2-D read-enabled memory and 2-D read-write memory, respectively. The application analysis shows that 2-D read-enabled memory achieves around 86% average decrease in row-buffer transactions in 256 × 256 size matrix operations without any array area increase. The 2-D read write memory offers 87% decrease in row-buffer transactions with 28.5% increase in array area compared to the 1-D FEFET memory.

AB - In this paper, we propose a novel memory architecture with the capability of single-cycle row-wise/column-wise accesses. Such an architecture is highly suitable for workloads featuring spatial locality in multiple dimensions, which is a characteristic of many matrix and array operations. We describe in detail the circuit design techniques enabling the proposed architectures, as well as the viability of emerging memory technologies based on ferroelectric transistors (FEFETs) for our design. Compared to FEFET memory with standard 1-D access, we achieve 5% energy savings for the proposed memory featuring 2-D read and 93% energy savings for memory with 2-D read and write, for 32 bit column read and write. In addition, we get around 11% and 95% delay savings for 2-D read-enabled memory and 2-D read-write memory, respectively. The application analysis shows that 2-D read-enabled memory achieves around 86% average decrease in row-buffer transactions in 256 × 256 size matrix operations without any array area increase. The 2-D read write memory offers 87% decrease in row-buffer transactions with 28.5% increase in array area compared to the 1-D FEFET memory.

UR - http://www.scopus.com/inward/record.url?scp=85042703281&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042703281&partnerID=8YFLogxK

U2 - 10.1109/TVLSI.2018.2801302

DO - 10.1109/TVLSI.2018.2801302

M3 - Article

AN - SCOPUS:85042703281

VL - 26

SP - 1040

EP - 1050

JO - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

JF - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

SN - 1063-8210

IS - 6

ER -