Performance comparison of some shared memory organizations for 2D mesh-like NOCs

Research output: Contribution to journalArticleScientificpeer-review

6 Citations (Scopus)

Abstract

While the research community has already studied a considerable amount of techniques related to achieving high bandwidth, good reliability, low power consumption, certain quality of service in communication on networks on chip (NOC) especially with artificial communication patterns, a little attention has paid to the effects of memory organizations to performance of computing engines employing NOCs with real parallel workloads. In this paper we compare the performance of some shared memory organizations for chip multiprocessors (CMP) employing advanced homogeneous 2D-mesh-like NOCs and making use of emulated shared memory and non-uniform memory access models. The evaluated techniques range from applying different hashing functions to elimination methods of speed difference between processing resources and memories, and from access methods to latency hiding and concurrent memory access support techniques. Tests are performed on our CMP/NOC framework with simple but real parallel programs that can be directly used as building blocks of larger explicitly parallel applications.
Original languageEnglish
Pages (from-to)274-284
Number of pages11
JournalMicroprocessors and Microsystems
Volume35
Issue number2
DOIs
Publication statusPublished - 2011
MoE publication typeA1 Journal article-refereed

Fingerprint

Data storage equipment
Communication
Quality of service
Electric power utilization
Engines
Bandwidth
Processing
Network-on-chip

Keywords

  • NOC
  • CMP
  • memory organization
  • hashing
  • speed difference elimination
  • latency hiding
  • concurrent access

Cite this

@article{00c7640f3d834487b7e8208630368b05,
title = "Performance comparison of some shared memory organizations for 2D mesh-like NOCs",
abstract = "While the research community has already studied a considerable amount of techniques related to achieving high bandwidth, good reliability, low power consumption, certain quality of service in communication on networks on chip (NOC) especially with artificial communication patterns, a little attention has paid to the effects of memory organizations to performance of computing engines employing NOCs with real parallel workloads. In this paper we compare the performance of some shared memory organizations for chip multiprocessors (CMP) employing advanced homogeneous 2D-mesh-like NOCs and making use of emulated shared memory and non-uniform memory access models. The evaluated techniques range from applying different hashing functions to elimination methods of speed difference between processing resources and memories, and from access methods to latency hiding and concurrent memory access support techniques. Tests are performed on our CMP/NOC framework with simple but real parallel programs that can be directly used as building blocks of larger explicitly parallel applications.",
keywords = "NOC, CMP, memory organization, hashing, speed difference elimination, latency hiding, concurrent access",
author = "Martti Forsell",
year = "2011",
doi = "10.1016/j.micpro.2010.07.003",
language = "English",
volume = "35",
pages = "274--284",
journal = "Microprocessors and Microsystems",
issn = "0141-9331",
publisher = "Elsevier",
number = "2",

}

Performance comparison of some shared memory organizations for 2D mesh-like NOCs. / Forsell, Martti.

In: Microprocessors and Microsystems, Vol. 35, No. 2, 2011, p. 274-284.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Performance comparison of some shared memory organizations for 2D mesh-like NOCs

AU - Forsell, Martti

PY - 2011

Y1 - 2011

N2 - While the research community has already studied a considerable amount of techniques related to achieving high bandwidth, good reliability, low power consumption, certain quality of service in communication on networks on chip (NOC) especially with artificial communication patterns, a little attention has paid to the effects of memory organizations to performance of computing engines employing NOCs with real parallel workloads. In this paper we compare the performance of some shared memory organizations for chip multiprocessors (CMP) employing advanced homogeneous 2D-mesh-like NOCs and making use of emulated shared memory and non-uniform memory access models. The evaluated techniques range from applying different hashing functions to elimination methods of speed difference between processing resources and memories, and from access methods to latency hiding and concurrent memory access support techniques. Tests are performed on our CMP/NOC framework with simple but real parallel programs that can be directly used as building blocks of larger explicitly parallel applications.

AB - While the research community has already studied a considerable amount of techniques related to achieving high bandwidth, good reliability, low power consumption, certain quality of service in communication on networks on chip (NOC) especially with artificial communication patterns, a little attention has paid to the effects of memory organizations to performance of computing engines employing NOCs with real parallel workloads. In this paper we compare the performance of some shared memory organizations for chip multiprocessors (CMP) employing advanced homogeneous 2D-mesh-like NOCs and making use of emulated shared memory and non-uniform memory access models. The evaluated techniques range from applying different hashing functions to elimination methods of speed difference between processing resources and memories, and from access methods to latency hiding and concurrent memory access support techniques. Tests are performed on our CMP/NOC framework with simple but real parallel programs that can be directly used as building blocks of larger explicitly parallel applications.

KW - NOC

KW - CMP

KW - memory organization

KW - hashing

KW - speed difference elimination

KW - latency hiding

KW - concurrent access

U2 - 10.1016/j.micpro.2010.07.003

DO - 10.1016/j.micpro.2010.07.003

M3 - Article

VL - 35

SP - 274

EP - 284

JO - Microprocessors and Microsystems

JF - Microprocessors and Microsystems

SN - 0141-9331

IS - 2

ER -