TY - GEN
T1 - An analysis of hpc benchmarks in virtual machine environments
AU - Tikotekar, Anand
AU - Vallée, Geoffroy
AU - Naughton, Thomas
AU - Ong, Hong
AU - Engelmann, Christian
AU - Scott, Stephen L.
PY - 2009
Y1 - 2009
N2 - Virtualization technology has been gaining acceptance in the scientific community due to its overall flexibility in running HPC applications. It has been reported that a specific class of applications is better suited to a particular type of virtualization scheme or implementation. For example, Xen has been shown to perform with little overhead for compute-bound applications. Such a study, although useful, does not allow us to generalize conclusions beyond the performance analysis of that application which is explicitly executed. An explanation of why the generalization described above is difficult, may be due to the versatility in applications, which leads to different overheads in virtual environments. For example, two similar applications may spend disproportionate amount of time in their respective library code when run in virtual environments. In this paper, we aim to study such potential causes by investigating the behavior and identifying patterns of various overheads for HPC benchmark applications. Based on the investigation of the overhead profiles for different benchmarks, we aim to address questions such as: Are the overhead profiles for a particular type of benchmarks (such as compute-bound) similar or are there grounds to conclude otherwise?
AB - Virtualization technology has been gaining acceptance in the scientific community due to its overall flexibility in running HPC applications. It has been reported that a specific class of applications is better suited to a particular type of virtualization scheme or implementation. For example, Xen has been shown to perform with little overhead for compute-bound applications. Such a study, although useful, does not allow us to generalize conclusions beyond the performance analysis of that application which is explicitly executed. An explanation of why the generalization described above is difficult, may be due to the versatility in applications, which leads to different overheads in virtual environments. For example, two similar applications may spend disproportionate amount of time in their respective library code when run in virtual environments. In this paper, we aim to study such potential causes by investigating the behavior and identifying patterns of various overheads for HPC benchmark applications. Based on the investigation of the overhead profiles for different benchmarks, we aim to address questions such as: Are the overhead profiles for a particular type of benchmarks (such as compute-bound) similar or are there grounds to conclude otherwise?
UR - http://www.scopus.com/inward/record.url?scp=68149165868&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-00955-6_8
DO - 10.1007/978-3-642-00955-6_8
M3 - Conference contribution
AN - SCOPUS:68149165868
SN - 3642009549
SN - 9783642009549
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 63
EP - 71
BT - Euro-Par 2008 Workshops - Parallel Processing - VHPC 2008, UNICORE 2008, HPPC 2008, SGS 2008, PROPER 2008, ROIA 2008, and DPA 2008, Revised Selected Papers
T2 - Workshops on Parallel Processing, Euro-Par 2008: VHPC 2008, UNICORE 2008, HPPC 2008, SGS 2008, PROPER 2008, ROIA 2008, and DPA 2008
Y2 - 25 August 2008 through 26 August 2008
ER -