Analytical performance models for NoCs with multiple priority traffic classes

Sumit K. Mandal, Raid Ayoub, Michael Kishinevsky, Umit Y. Ogras

Research output: Contribution to journalArticle

Abstract

Networks-on-chip (NoCs) have become the standard for interconnect solutions in industrial designs ranging from client CPUs to many-core chip-multiprocessors. Since NoCs play a vital role in system performance and power consumption, pre-silicon evaluation environments include cycle-accurate NoC simulators. Long simulations increase the execution time of evaluation frameworks, which are already notoriously slow, and prohibit design-space exploration. Existing analytical NoC models, which assume fair arbitration, cannot replace these simulations since industrial NoCs typically employ priority schedulers and multiple priority classes. To address this limitation, we propose a systematic approach to construct priority-aware analytical performance models using micro-architecture specifications and input traffic. Our approach decomposes the given NoC into individual queues with modified service time to enable accurate and scalable latency computations. Specifically, we introduce novel transformations along with an algorithm that iteratively applies these transformations to decompose the queuing system. Experimental evaluations using real architectures and applications show high accuracy of 97% and up to 2.5× speedup in full-system simulation.

Original languageEnglish (US)
Article numbera52
JournalACM Transactions on Embedded Computing Systems
Volume18
Issue number5s
DOIs
StatePublished - Oct 2019

Fingerprint

Product design
Program processors
Electric power utilization
Simulators
Specifications
Silicon
Network-on-chip

Keywords

  • NoC performance analysis
  • Priority-based NoC
  • Queuing networks

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture

Cite this

Analytical performance models for NoCs with multiple priority traffic classes. / Mandal, Sumit K.; Ayoub, Raid; Kishinevsky, Michael; Ogras, Umit Y.

In: ACM Transactions on Embedded Computing Systems, Vol. 18, No. 5s, a52, 10.2019.

Research output: Contribution to journalArticle

@article{ff166f60cc6c47d5a91d938496f0ded0,
title = "Analytical performance models for NoCs with multiple priority traffic classes",
abstract = "Networks-on-chip (NoCs) have become the standard for interconnect solutions in industrial designs ranging from client CPUs to many-core chip-multiprocessors. Since NoCs play a vital role in system performance and power consumption, pre-silicon evaluation environments include cycle-accurate NoC simulators. Long simulations increase the execution time of evaluation frameworks, which are already notoriously slow, and prohibit design-space exploration. Existing analytical NoC models, which assume fair arbitration, cannot replace these simulations since industrial NoCs typically employ priority schedulers and multiple priority classes. To address this limitation, we propose a systematic approach to construct priority-aware analytical performance models using micro-architecture specifications and input traffic. Our approach decomposes the given NoC into individual queues with modified service time to enable accurate and scalable latency computations. Specifically, we introduce novel transformations along with an algorithm that iteratively applies these transformations to decompose the queuing system. Experimental evaluations using real architectures and applications show high accuracy of 97{\%} and up to 2.5× speedup in full-system simulation.",
keywords = "NoC performance analysis, Priority-based NoC, Queuing networks",
author = "Mandal, {Sumit K.} and Raid Ayoub and Michael Kishinevsky and Ogras, {Umit Y.}",
year = "2019",
month = "10",
doi = "10.1145/3358176",
language = "English (US)",
volume = "18",
journal = "ACM Transactions on Embedded Computing Systems",
issn = "1539-9087",
publisher = "Association for Computing Machinery (ACM)",
number = "5s",

}

TY - JOUR

T1 - Analytical performance models for NoCs with multiple priority traffic classes

AU - Mandal, Sumit K.

AU - Ayoub, Raid

AU - Kishinevsky, Michael

AU - Ogras, Umit Y.

PY - 2019/10

Y1 - 2019/10

N2 - Networks-on-chip (NoCs) have become the standard for interconnect solutions in industrial designs ranging from client CPUs to many-core chip-multiprocessors. Since NoCs play a vital role in system performance and power consumption, pre-silicon evaluation environments include cycle-accurate NoC simulators. Long simulations increase the execution time of evaluation frameworks, which are already notoriously slow, and prohibit design-space exploration. Existing analytical NoC models, which assume fair arbitration, cannot replace these simulations since industrial NoCs typically employ priority schedulers and multiple priority classes. To address this limitation, we propose a systematic approach to construct priority-aware analytical performance models using micro-architecture specifications and input traffic. Our approach decomposes the given NoC into individual queues with modified service time to enable accurate and scalable latency computations. Specifically, we introduce novel transformations along with an algorithm that iteratively applies these transformations to decompose the queuing system. Experimental evaluations using real architectures and applications show high accuracy of 97% and up to 2.5× speedup in full-system simulation.

AB - Networks-on-chip (NoCs) have become the standard for interconnect solutions in industrial designs ranging from client CPUs to many-core chip-multiprocessors. Since NoCs play a vital role in system performance and power consumption, pre-silicon evaluation environments include cycle-accurate NoC simulators. Long simulations increase the execution time of evaluation frameworks, which are already notoriously slow, and prohibit design-space exploration. Existing analytical NoC models, which assume fair arbitration, cannot replace these simulations since industrial NoCs typically employ priority schedulers and multiple priority classes. To address this limitation, we propose a systematic approach to construct priority-aware analytical performance models using micro-architecture specifications and input traffic. Our approach decomposes the given NoC into individual queues with modified service time to enable accurate and scalable latency computations. Specifically, we introduce novel transformations along with an algorithm that iteratively applies these transformations to decompose the queuing system. Experimental evaluations using real architectures and applications show high accuracy of 97% and up to 2.5× speedup in full-system simulation.

KW - NoC performance analysis

KW - Priority-based NoC

KW - Queuing networks

UR - http://www.scopus.com/inward/record.url?scp=85073152155&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073152155&partnerID=8YFLogxK

U2 - 10.1145/3358176

DO - 10.1145/3358176

M3 - Article

AN - SCOPUS:85073152155

VL - 18

JO - ACM Transactions on Embedded Computing Systems

JF - ACM Transactions on Embedded Computing Systems

SN - 1539-9087

IS - 5s

M1 - a52

ER -