CFD Solvers with Minimal Memory Access

Many state of the art CFD codes that exhibit low computational intensity (flops per RAM access) "saturate" the memory bandwidth of modern chips after only a few cores, thus minimizing any benefits from going to a higher number of available cores. This bottleneck is expected to become even...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autor principal: Löhner, Rainald
Formato: Objeto de conferencia Resumen
Lenguaje:Inglés
Publicado: 2017
Materias:
Ingeniería
CFD codes
Memory bandwidth
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/94120
https://cimec.org.ar/ojs/index.php/mc/article/view/5230
Aporte de:
SEDICI (UNLP) de Universidad Nacional de La Plata
id I19-R120-10915-94120
record_format dspace
institution Universidad Nacional de La Plata
institution_str I-19
repository_str R-120
collection SEDICI (UNLP)
language Inglés
topic Ingeniería
CFD codes
Memory bandwidth
spellingShingle Ingeniería
CFD codes
Memory bandwidth
Löhner, Rainald
CFD Solvers with Minimal Memory Access
topic_facet Ingeniería
CFD codes
Memory bandwidth
description Many state of the art CFD codes that exhibit low computational intensity (flops per RAM access) "saturate" the memory bandwidth of modern chips after only a few cores, thus minimizing any benefits from going to a higher number of available cores. This bottleneck is expected to become even more pronounced for future manycore systems. This has led to the quest for CFD solvers with minimal memory access. We report on recent developments and results for Finite Difference and Edge-Based Finite Element solvers. The best of these implementations yield one residual for only 6 fetches and 4 stores, regardless of the size of the stencil (and therefore the discretization order). This means that in terms of memory access they are competitive even with finite difference stencils as low as 2 (typical of CFD codes with 2nd order spatial discretization of fluxes and 4th order damping). Timings for a low Mach number finite difference code using a 6th order spatial discretization show competitive timings as compared to conventional loops. This bodes well for future HPC architectures.
format Objeto de conferencia
Resumen
author Löhner, Rainald
author_facet Löhner, Rainald
author_sort Löhner, Rainald
title CFD Solvers with Minimal Memory Access
title_short CFD Solvers with Minimal Memory Access
title_full CFD Solvers with Minimal Memory Access
title_fullStr CFD Solvers with Minimal Memory Access
title_full_unstemmed CFD Solvers with Minimal Memory Access
title_sort cfd solvers with minimal memory access
publishDate 2017
url http://sedici.unlp.edu.ar/handle/10915/94120
https://cimec.org.ar/ojs/index.php/mc/article/view/5230
work_keys_str_mv AT lohnerrainald cfdsolverswithminimalmemoryaccess
bdutipo_str Repositorios
_version_ 1764820492142772229

Ejemplares similares