A high-performance portable abstract interface for explicit SIMD vectorization

This work establishes a scalable, easy to use and efficient approach for exploiting SIMD capabilities of modern CPUs, without the need for extensive knowledge of architecture specific instruction sets. We provide a description of a new API, known as UME::SIMD, which provides a flexible, portable, ty...

Descripción completa

Detalles Bibliográficos
Autores principales: Karpiński, P, McDonald, J
Lenguaje:eng
Publicado: 2017
Materias:
Computing and Computers
Acceso en línea:https://dx.doi.org/10.1145/3026937.3026939
http://cds.cern.ch/record/2318246
_version_ 1780958448969580544
author Karpiński, P
McDonald, J
author_facet Karpiński, P
McDonald, J
author_sort Karpiński, P
collection CERN
description This work establishes a scalable, easy to use and efficient approach for exploiting SIMD capabilities of modern CPUs, without the need for extensive knowledge of architecture specific instruction sets. We provide a description of a new API, known as UME::SIMD, which provides a flexible, portable, type-oriented abstraction for SIMD instruction set architectures. Requirements for such libraries are analysed based on existing, as well as proposed future solutions. A software architecture that achieves these requirements is explained, and its performance evaluated. Finally we discuss how the API fits into the existing, and future software ecosystem.
id oai-inspirehep.net-1670543
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2017
record_format invenio
spelling oai-inspirehep.net-16705432019-09-30T06:29:59Zdoi:10.1145/3026937.3026939http://cds.cern.ch/record/2318246engKarpiński, PMcDonald, JA high-performance portable abstract interface for explicit SIMD vectorizationComputing and ComputersThis work establishes a scalable, easy to use and efficient approach for exploiting SIMD capabilities of modern CPUs, without the need for extensive knowledge of architecture specific instruction sets. We provide a description of a new API, known as UME::SIMD, which provides a flexible, portable, type-oriented abstraction for SIMD instruction set architectures. Requirements for such libraries are analysed based on existing, as well as proposed future solutions. A software architecture that achieves these requirements is explained, and its performance evaluated. Finally we discuss how the API fits into the existing, and future software ecosystem.oai:inspirehep.net:16705432017
spellingShingle Computing and Computers
Karpiński, P
McDonald, J
A high-performance portable abstract interface for explicit SIMD vectorization
title A high-performance portable abstract interface for explicit SIMD vectorization
title_full A high-performance portable abstract interface for explicit SIMD vectorization
title_fullStr A high-performance portable abstract interface for explicit SIMD vectorization
title_full_unstemmed A high-performance portable abstract interface for explicit SIMD vectorization
title_short A high-performance portable abstract interface for explicit SIMD vectorization
title_sort high-performance portable abstract interface for explicit simd vectorization
topic Computing and Computers
url https://dx.doi.org/10.1145/3026937.3026939
http://cds.cern.ch/record/2318246
work_keys_str_mv AT karpinskip ahighperformanceportableabstractinterfaceforexplicitsimdvectorization
AT mcdonaldj ahighperformanceportableabstractinterfaceforexplicitsimdvectorization
AT karpinskip highperformanceportableabstractinterfaceforexplicitsimdvectorization
AT mcdonaldj highperformanceportableabstractinterfaceforexplicitsimdvectorization

Ejemplares similares