Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application
Identifieur interne : 002971 ( Crin/Curation ); précédent : 002970; suivant : 002972Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application
Auteurs : Xavier Cavin ; Laurent AlonsoSource :
English descriptors
Abstract
We show in this paper the memory management issues raised by a parallel irregular and dynamic hierarchical application, which constantly allocates and deallocates data over an extremely large virtual address space. First, we show that if memory caches data locality is necessary, a lack of virtual pages locality may greatly affect the obtained performance. Second, fragmentation and contention problems associated with the required parallel dynamic memory allocation are presented. We propose practical solutions and discuss experimentation results obtained on a cache-coherent non uniform memory access (ccNUMA) distributed shared memory SGI Origin2000 machine.
Links toward previous steps (curation, corpus...)
- to stream Crin, to step Corpus: Pour aller vers cette notice dans l'étape Curation :002971
Links to Exploration step
CRIN:cavin00aLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="464">Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application</title>
</titleStmt>
<publicationStmt><idno type="RBID">CRIN:cavin00a</idno>
<date when="2000" year="2000">2000</date>
<idno type="wicri:Area/Crin/Corpus">002971</idno>
<idno type="wicri:Area/Crin/Curation">002971</idno>
<idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Curation">002971</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en">Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application</title>
<author><name sortKey="Cavin, Xavier" sort="Cavin, Xavier" uniqKey="Cavin X" first="Xavier" last="Cavin">Xavier Cavin</name>
</author>
<author><name sortKey="Alonso, Laurent" sort="Alonso, Laurent" uniqKey="Alonso L" first="Laurent" last="Alonso">Laurent Alonso</name>
</author>
</analytic>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>dynamic allocation</term>
<term>fem</term>
<term>hierarchical algorithm</term>
<term>memory management</term>
<term>parallelism</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en" wicri:score="2061">We show in this paper the memory management issues raised by a parallel irregular and dynamic hierarchical application, which constantly allocates and deallocates data over an extremely large virtual address space. First, we show that if memory caches data locality is necessary, a lack of virtual pages locality may greatly affect the obtained performance. Second, fragmentation and contention problems associated with the required parallel dynamic memory allocation are presented. We propose practical solutions and discuss experimentation results obtained on a cache-coherent non uniform memory access (ccNUMA) distributed shared memory SGI Origin2000 machine.</div>
</front>
</TEI>
<BibTex type="inproceedings"><ref>cavin00a</ref>
<crinnumber>A00-R-096</crinnumber>
<category>3</category>
<equipe>ISA</equipe>
<author><e>Cavin, Xavier</e>
<e>Alonso, Laurent</e>
</author>
<title>Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application</title>
<booktitle>{Seventh International Workshop On Solving Irregularly Structured Problems In Parallel - Irregular'2000, Cancun, Mexico}</booktitle>
<year>2000</year>
<editor>Jose Rolim et al.</editor>
<volume>1800</volume>
<series>Lecture Notes in Computer Science</series>
<pages>428-434</pages>
<month>May</month>
<organization>S. Sahni, T. Davis, S. Rajasekeran, S. Ranka</organization>
<publisher>Springer</publisher>
<keywords><e>memory management</e>
<e>dynamic allocation</e>
<e>parallelism</e>
<e>hierarchical algorithm</e>
<e>fem</e>
</keywords>
<abstract>We show in this paper the memory management issues raised by a parallel irregular and dynamic hierarchical application, which constantly allocates and deallocates data over an extremely large virtual address space. First, we show that if memory caches data locality is necessary, a lack of virtual pages locality may greatly affect the obtained performance. Second, fragmentation and contention problems associated with the required parallel dynamic memory allocation are presented. We propose practical solutions and discuss experimentation results obtained on a cache-coherent non uniform memory access (ccNUMA) distributed shared memory SGI Origin2000 machine.</abstract>
</BibTex>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Crin/Curation
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 002971 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Crin/Curation/biblio.hfd -nk 002971 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Lorraine |area= InforLorV4 |flux= Crin |étape= Curation |type= RBID |clé= CRIN:cavin00a |texte= Parallel Management of Large Dynamic Shared Memory Space : a Hierarchical FEM Application }}
This area was generated with Dilib version V0.6.33. |