Parallel bottleneck in the Quasineutrality solver embedded in GYSELA

نویسندگان

  • G. Latu
  • V. Grandgirard
چکیده

This report shows some performance results of the Quasineutraliy Poisson solver used in the GYSELA code. The numerical schemes of this Poisson solver is explained, and the computation and communication steps on a parallel machine are described. Benchmarks shows several time measurement from 32 cores to 4096 cores. Present bottlenecks and problems of the parallel algorithm are discussed. Some possible solutions are foreseen. Key-words: Quasineutrality solver, Gyrokinetics, MPI ∗ CEA Cadarache, 13108 Saint-Paul-les-Durance Cedex † INRIA Nancy-Grand Est & Université de Strasbourg, 7 rue Descartes, 6700 Strasbourg in ria -0 05 83 68 9, v er si on 1 6 Ap r 2 01 1 Limitations associées à la parallélisation du solveur Quasi-neutre inclus dans GYSELA Résumé : Ce rapport présente des prises de performances du solveur Poisson Quasi-neutre utilisé dans le code GYSELA. Le schéma numérique de ce solveur Poisson est décrit, ainsi que les di érentes étapes de calculs et de communication sur machine parallèle. Une série de benchmarks on été e ectués de 32 à 4096 c÷urs, cela donne un aperçu des performances de ce solveur parallèle. Les goulots d'étranglement et les limitations de l'algorithme parallèle utilisé sont explicités. En n, des solutions possibles sont envisagées. Mots-clés : Solveur Quasi-neutre, Gyrocinétique, MPI in ria -0 05 83 68 9, v er si on 1 6 Ap r 2 01 1 Parallel bottleneck of GYSELA's Poisson solver 3

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Scalable Quasineutral Solver for Gyrokinetic Simulation

Modeling turbulent transport is a major goal in order to predict con nement issues in a tokamak plasma. The gyrokinetic framework considers a computational domain in ve dimensions to look at kinetic issues in a plasma. Gyrokinetic simulations lead to huge computational needs. Up to now, the gyrokinetic code GYSELA performed large simulations using a few thousands of cores. The work proposed her...

متن کامل

Heuristic approach to solve hybrid flow shop scheduling problem with unrelated parallel machines

In hybrid flow shop scheduling problem (HFS) with unrelated parallel machines, a set of n jobs are processed on k machines. A mixed integer linear programming (MILP) model for the HFS scheduling problems with unrelated parallel machines has been proposed to minimize the maximum completion time (makespan). Since the problem is shown to be NP-complete, it is necessary to use heuristic methods to ...

متن کامل

A Message-Passing Distributed Memory Parallel Algorithm for a Dual-Code Thin Layer, Parabolized Navier-Stokes Solver

In this study, the results of parallelization of a 3-D dual code (Thin Layer, Parabolized Navier-Stokes solver) for solving supersonic turbulent flow around body and wing-body combinations are presented. As a serial code, TLNS solver is very time consuming and takes a large part of memory due to the iterative and lengthy computations. Also for complicated geometries, an exceeding number of grid...

متن کامل

A High Performance Parallel IP Lookup Technique Using Distributed Memory Organization and ISCB-Tree Data Structure

The IP Lookup Process is a key bottleneck in routing due to the increase in routing table size, increasing traıc and migration to IPv6 addresses. The IP address lookup involves computation of the Longest Prefix Matching (LPM), which existing solutions such as BSD Radix Tries, scale poorly when traıc in the router increases or when employed for IPv6 address lookups. In this paper, we describe a ...

متن کامل

A High Performance Parallel IP Lookup Technique Using Distributed Memory Organization and ISCB-Tree Data Structure

The IP Lookup Process is a key bottleneck in routing due to the increase in routing table size, increasing traıc and migration to IPv6 addresses. The IP address lookup involves computation of the Longest Prefix Matching (LPM), which existing solutions such as BSD Radix Tries, scale poorly when traıc in the router increases or when employed for IPv6 address lookups. In this paper, we describe a ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2011