Mostra el registre d'ítem simple
Using Arm’s scalable vector extension on stencil codes
dc.contributor.author | Armejach Sanosa, Adrià |
dc.contributor.author | Caminal Pallarés, Helena |
dc.contributor.author | Cebrián González, Juan Manuel |
dc.contributor.author | Langarita, Rubén |
dc.contributor.author | González-Alberquilla, Rekai |
dc.contributor.author | Adeniyi-Jones, Chris |
dc.contributor.author | Valero Cortés, Mateo |
dc.contributor.author | Casas, Marc |
dc.contributor.author | Moretó Planas, Miquel |
dc.contributor.other | Universitat Politècnica de Catalunya. Departament d'Arquitectura de Computadors |
dc.date.accessioned | 2019-10-08T07:00:16Z |
dc.date.available | 2020-04-08T00:25:58Z |
dc.date.issued | 2020-03 |
dc.identifier.citation | Armejach, A. [et al.]. Using Arm’s scalable vector extension on stencil codes. "Journal of supercomputing", vol. 76, Març 2020, p. 2039-2062. |
dc.identifier.issn | 0920-8542 |
dc.identifier.uri | http://hdl.handle.net/2117/169340 |
dc.description.abstract | Data-level parallelism is frequently ignored or underutilized. Achieved through vector/SIMD capabilities, it can provide substantial performance improvements on top of widely used techniques such as thread-level parallelism. However, manual vectorization is a tedious and costly process that needs to be repeated for each specific instruction set or register size. In addition, automatic compiler vectorization is susceptible to code complexity, and usually limited due to data and control dependencies. To address some of these issues, Arm recently released a new vector ISA, the scalable vector extension (SVE), which is vector-length agnostic (VLA). VLA enables the generation of binary files that run regardless of the physical vector register length. In this paper, we leverage the main characteristics of SVE to implement and optimize stencil computations, ubiquitous in scientific computing. We show that SVE enables easy deployment of textbook optimizations like loop unrolling, loop fusion, load trading or data reuse. Our detailed simulations using vector lengths ranging from 128 to 2048 bits show that these optimizations can lead to performance improvements over straightforward vectorized code of up to 1.57×. In addition, we show that certain optimizations can hurt performance due to reduced arithmetic intensity and instruction overheads, and provide insight useful for compiler optimizers. |
dc.format.extent | 24 p. |
dc.language.iso | eng |
dc.subject | Àrees temàtiques de la UPC::Informàtica::Arquitectura de computadors |
dc.subject.lcsh | Ubiquitous computing |
dc.subject.lcsh | Compilers (Computer programs) |
dc.subject.other | Data-level parallelism |
dc.subject.other | Scalable vector extension |
dc.subject.other | Vector-length agnostic |
dc.subject.other | Stencil computations |
dc.title | Using Arm’s scalable vector extension on stencil codes |
dc.type | Article |
dc.subject.lemac | Informàtica ubiqua |
dc.subject.lemac | Compiladors (Programes d'ordinador) |
dc.contributor.group | Universitat Politècnica de Catalunya. CAP - Grup de Computació d'Altes Prestacions |
dc.identifier.doi | 10.1007/s11227-019-02842-5 |
dc.description.peerreviewed | Peer Reviewed |
dc.relation.publisherversion | https://link.springer.com/article/10.1007/s11227-019-02842-5 |
dc.rights.access | Open Access |
local.identifier.drac | 25169572 |
dc.description.version | Postprint (author's final draft) |
dc.relation.projectid | info:eu-repo/grantAgreement/AEI/RYC-2016-21104 |
dc.relation.projectid | info:eu-repo/grantAgreement/AGAUR/2017 SGR 1414 |
dc.relation.projectid | info:eu-repo/grantAgreement/MINECO//TIN2015-65316-P/ES/COMPUTACION DE ALTAS PRESTACIONES VII/ |
local.citation.author | Armejach, A.; Caminal, H.; Cebrián, J. M.; Langarita, R.; González-Alberquilla, R.; Adeniyi-Jones, C.; Valero, M.; Casas, M.; Moreto, M. |
local.citation.publicationName | Journal of supercomputing |
local.citation.volume | 76 |
local.citation.startingPage | 2039 |
local.citation.endingPage | 2062 |
Fitxers d'aquest items
Aquest ítem apareix a les col·leccions següents
-
Articles de revista [1.049]
-
Articles de revista [382]