eprintid: 11585
rev_number: 7
eprint_status: archive
userid: 1
dir: disk0/00/01/15/85
datestamp: 2011-02-03 15:51:55
lastmod: 2014-08-19 02:25:46
status_changed: 2012-08-15 08:52:34
type: article
metadata_visibility: show
creators_name: Mücke, Manfred
creators_name: Gansterer, Wilfried N.
title: Analysis of Inter-Chip Communication Patterns on Multi-Core Distributed Shared-Memory Computers
ispublished: pub
subjects: 004
divisions: 720000
cterms_swd: Computerarchitektur
abstract: Multi-core multi-socket distributed shared-memory com- puters (DSM computers, for short) have become an impor- tant node architecture in scientific computing as they provide substantial computational capacity with relatively low space and power requirements. Compared to conventional computer networks, inter-chip networks used in DSM computers feature higher bandwidth, lower latency and tighter integration with the CPU. The inter-chip network is a shared resource among the user application and many other services, which can lead to consid- erable variation of execution times of identical communication tasks. In this work, we explore traffic patterns resulting from MPI collective communication primitives and investigate the ques- tion whether inter-chip link load is a reliable indicator and predictor for the execution time of collective communication primitives on a DSM computer. Our experiments on a Sun Fire X4600 M2 DSM computer with 32 cores (eight quad-core CPUs) indicate that specific single link loads are positively correlated with the execution time of MPI ALLREDUCE. Ob- serving patterns over multiple links allows refinement of the single-link observation.
abstract_translated_lang: eng
date: 2011
date_type: published
id_scheme: DOI
id_number: 10.11588/heidok.00011585
portal_cluster_id: p-whtra11
portal_order: WHTRA2011
ppn_swb: 1407033808
own_urn: urn:nbn:de:bsz:16-opus-115852
language: eng
bibsort: MUCKEMANFRANALYSISOF2011
full_text_status: public
citation:   Mücke, Manfred ; Gansterer, Wilfried N.  (2011) Analysis of Inter-Chip Communication Patterns on Multi-Core Distributed Shared-Memory Computers.         
document_url: https://archiv.ub.uni-heidelberg.de/volltextserver/11585/1/WHTRA2011_paper01.pdf