TCASM

An asynchronous shared memory interface for high-performance application composition

Douglas Otstott, Latchesar Ionkov, Michael Lang, Ming Zhao

Research output: Contribution to journalArticle

Abstract

This paper addresses the growing need for mechanisms supporting intra-node application composition in high-performance computing (HPC) systems. It provides a novel shared memory interface that allows composite applications, two or more coupled applications, to share internal data structures without blocking. This allows independent progress of the applications such that they can proceed in a parallel, overlapped fashion. Composite applications using in-node shared memory can reduce the amount of data to be communicated between nodes, allowing checkpointing and data reduction or analytics to be performed locally and in parallel. The approach is implemented in Linux, and evaluated using benchmarks that represent typical composite applications on a large HPC testbed. The results show that the proposed approach significantly outperforms the traditional ones (up to a 15-fold speed increase on a 200 node machine).

Original languageEnglish (US)
JournalParallel Computing
DOIs
StateAccepted/In press - Mar 22 2016

Fingerprint

Shared Memory
Interfaces (computer)
High Performance
Data storage equipment
Chemical analysis
Composite
Vertex of a graph
Composite materials
Checkpointing
Computing
Data Reduction
Linux
Testbeds
Testbed
Data structures
Data reduction
Data Structures
Fold
Computer systems
Benchmark

Keywords

  • Checkpointing
  • Composite applications
  • In situ analysis
  • Shared memory

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Software
  • Hardware and Architecture
  • Computer Networks and Communications
  • Computer Graphics and Computer-Aided Design
  • Artificial Intelligence

Cite this

TCASM : An asynchronous shared memory interface for high-performance application composition. / Otstott, Douglas; Ionkov, Latchesar; Lang, Michael; Zhao, Ming.

In: Parallel Computing, 22.03.2016.

Research output: Contribution to journalArticle

@article{b0043587bc1348bab999be681c3ace4b,
title = "TCASM: An asynchronous shared memory interface for high-performance application composition",
abstract = "This paper addresses the growing need for mechanisms supporting intra-node application composition in high-performance computing (HPC) systems. It provides a novel shared memory interface that allows composite applications, two or more coupled applications, to share internal data structures without blocking. This allows independent progress of the applications such that they can proceed in a parallel, overlapped fashion. Composite applications using in-node shared memory can reduce the amount of data to be communicated between nodes, allowing checkpointing and data reduction or analytics to be performed locally and in parallel. The approach is implemented in Linux, and evaluated using benchmarks that represent typical composite applications on a large HPC testbed. The results show that the proposed approach significantly outperforms the traditional ones (up to a 15-fold speed increase on a 200 node machine).",
keywords = "Checkpointing, Composite applications, In situ analysis, Shared memory",
author = "Douglas Otstott and Latchesar Ionkov and Michael Lang and Ming Zhao",
year = "2016",
month = "3",
day = "22",
doi = "10.1016/j.parco.2017.01.003",
language = "English (US)",
journal = "Parallel Computing",
issn = "0167-8191",
publisher = "Elsevier",

}

TY - JOUR

T1 - TCASM

T2 - An asynchronous shared memory interface for high-performance application composition

AU - Otstott, Douglas

AU - Ionkov, Latchesar

AU - Lang, Michael

AU - Zhao, Ming

PY - 2016/3/22

Y1 - 2016/3/22

N2 - This paper addresses the growing need for mechanisms supporting intra-node application composition in high-performance computing (HPC) systems. It provides a novel shared memory interface that allows composite applications, two or more coupled applications, to share internal data structures without blocking. This allows independent progress of the applications such that they can proceed in a parallel, overlapped fashion. Composite applications using in-node shared memory can reduce the amount of data to be communicated between nodes, allowing checkpointing and data reduction or analytics to be performed locally and in parallel. The approach is implemented in Linux, and evaluated using benchmarks that represent typical composite applications on a large HPC testbed. The results show that the proposed approach significantly outperforms the traditional ones (up to a 15-fold speed increase on a 200 node machine).

AB - This paper addresses the growing need for mechanisms supporting intra-node application composition in high-performance computing (HPC) systems. It provides a novel shared memory interface that allows composite applications, two or more coupled applications, to share internal data structures without blocking. This allows independent progress of the applications such that they can proceed in a parallel, overlapped fashion. Composite applications using in-node shared memory can reduce the amount of data to be communicated between nodes, allowing checkpointing and data reduction or analytics to be performed locally and in parallel. The approach is implemented in Linux, and evaluated using benchmarks that represent typical composite applications on a large HPC testbed. The results show that the proposed approach significantly outperforms the traditional ones (up to a 15-fold speed increase on a 200 node machine).

KW - Checkpointing

KW - Composite applications

KW - In situ analysis

KW - Shared memory

UR - http://www.scopus.com/inward/record.url?scp=85017100853&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85017100853&partnerID=8YFLogxK

U2 - 10.1016/j.parco.2017.01.003

DO - 10.1016/j.parco.2017.01.003

M3 - Article

JO - Parallel Computing

JF - Parallel Computing

SN - 0167-8191

ER -