The Bitlet Model: A Parameterized Analytical Model to Compare PIM and CPU Systems

Ronny Ronen; Adi Eliahu; Orian Leitersdorf; Natan Peled; Kunal Korgaonkar; Anupam Chattopadhyay; Ben Perach; Shahar Kvatinsky

doi:10.1145/3465371

The Bitlet Model: A Parameterized Analytical Model to Compare PIM and CPU Systems

Ronny Ronen, Adi Eliahu, Orian Leitersdorf, Natan Peled, Kunal Korgaonkar, Anupam Chattopadhyay, Ben Perach, Shahar Kvatinsky

Electrical and Computer Engineering

Technion - Israel Institute of Technology

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

Abstract

Currently, data-intensive applications are gaining popularity. Together with this trend, processing-in-memory (PIM)-based systems are being given more attention and have become more relevant. This article describes an analytical modeling tool called Bitlet that can be used in a parameterized fashion to estimate the performance and power/energy of a PIM-based system and, thereby, assess the affinity of workloads for PIM as opposed to traditional computing. The tool uncovers interesting trade-offs between, mainly, the PIM computation complexity (cycles required to perform a computation through PIM), the amount of memory used for PIM, the system memory bandwidth, and the data transfer size. Despite its simplicity, the model reveals new insights when applied to real-life examples. The model is demonstrated for several synthetic examples and then applied to explore the influence of different parameters on two systems - IMAGING and FloatPIM. Based on the demonstrations, insights about PIM and its combination with a CPU are provided.

Original language	English
Article number	43
Journal	ACM Journal on Emerging Technologies in Computing Systems
Volume	18
Issue number	2
DOIs	https://doi.org/10.1145/3465371
State	Published - Apr 2022

Keywords

Memristive memory
analytical models
non-volatile memory
processing in memory

All Science Journal Classification (ASJC) codes

Software
Hardware and Architecture
Electrical and Electronic Engineering

Access to Document

10.1145/3465371

Cite this

@article{9797d8a6802b46af87285b1ba511c4a9,

title = "The Bitlet Model: A Parameterized Analytical Model to Compare PIM and CPU Systems",

abstract = "Currently, data-intensive applications are gaining popularity. Together with this trend, processing-in-memory (PIM)-based systems are being given more attention and have become more relevant. This article describes an analytical modeling tool called Bitlet that can be used in a parameterized fashion to estimate the performance and power/energy of a PIM-based system and, thereby, assess the affinity of workloads for PIM as opposed to traditional computing. The tool uncovers interesting trade-offs between, mainly, the PIM computation complexity (cycles required to perform a computation through PIM), the amount of memory used for PIM, the system memory bandwidth, and the data transfer size. Despite its simplicity, the model reveals new insights when applied to real-life examples. The model is demonstrated for several synthetic examples and then applied to explore the influence of different parameters on two systems - IMAGING and FloatPIM. Based on the demonstrations, insights about PIM and its combination with a CPU are provided.",

keywords = "Memristive memory, analytical models, non-volatile memory, processing in memory",

author = "Ronny Ronen and Adi Eliahu and Orian Leitersdorf and Natan Peled and Kunal Korgaonkar and Anupam Chattopadhyay and Ben Perach and Shahar Kvatinsky",

note = "Publisher Copyright: {\textcopyright} 2022 Association for Computing Machinery.",

year = "2022",

month = apr,

doi = "10.1145/3465371",

language = "الإنجليزيّة",

volume = "18",

journal = "ACM Journal on Emerging Technologies in Computing Systems",

issn = "1550-4832",

publisher = "Association for Computing Machinery (ACM)",

number = "2",

}

TY - JOUR

T1 - The Bitlet Model

T2 - A Parameterized Analytical Model to Compare PIM and CPU Systems

AU - Ronen, Ronny

AU - Eliahu, Adi

AU - Leitersdorf, Orian

AU - Peled, Natan

AU - Korgaonkar, Kunal

AU - Chattopadhyay, Anupam

AU - Perach, Ben

AU - Kvatinsky, Shahar

PY - 2022/4

Y1 - 2022/4

N2 - Currently, data-intensive applications are gaining popularity. Together with this trend, processing-in-memory (PIM)-based systems are being given more attention and have become more relevant. This article describes an analytical modeling tool called Bitlet that can be used in a parameterized fashion to estimate the performance and power/energy of a PIM-based system and, thereby, assess the affinity of workloads for PIM as opposed to traditional computing. The tool uncovers interesting trade-offs between, mainly, the PIM computation complexity (cycles required to perform a computation through PIM), the amount of memory used for PIM, the system memory bandwidth, and the data transfer size. Despite its simplicity, the model reveals new insights when applied to real-life examples. The model is demonstrated for several synthetic examples and then applied to explore the influence of different parameters on two systems - IMAGING and FloatPIM. Based on the demonstrations, insights about PIM and its combination with a CPU are provided.

AB - Currently, data-intensive applications are gaining popularity. Together with this trend, processing-in-memory (PIM)-based systems are being given more attention and have become more relevant. This article describes an analytical modeling tool called Bitlet that can be used in a parameterized fashion to estimate the performance and power/energy of a PIM-based system and, thereby, assess the affinity of workloads for PIM as opposed to traditional computing. The tool uncovers interesting trade-offs between, mainly, the PIM computation complexity (cycles required to perform a computation through PIM), the amount of memory used for PIM, the system memory bandwidth, and the data transfer size. Despite its simplicity, the model reveals new insights when applied to real-life examples. The model is demonstrated for several synthetic examples and then applied to explore the influence of different parameters on two systems - IMAGING and FloatPIM. Based on the demonstrations, insights about PIM and its combination with a CPU are provided.

KW - Memristive memory

KW - analytical models

KW - non-volatile memory

KW - processing in memory

UR - http://www.scopus.com/inward/record.url?scp=85129742863&partnerID=8YFLogxK

U2 - 10.1145/3465371

DO - 10.1145/3465371

M3 - مقالة

SN - 1550-4832

VL - 18

JO - ACM Journal on Emerging Technologies in Computing Systems

JF - ACM Journal on Emerging Technologies in Computing Systems

IS - 2

M1 - 43

ER -

The Bitlet Model: A Parameterized Analytical Model to Compare PIM and CPU Systems

Abstract

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this