[Omp] performance study of spec.omp

[Karl Fuerlinger]

Hi Soohong,

I have recently done a study of the SPEC OpenMP benchmarks and their 
overhead and scalability characteristis on SGI Altix machines. The paper 
has been accepted for presentation at ICCS 2007. I've put it online here, 
maybe it is interesting to you:

http://www.cs.utk.edu/~karl/research/pubs/FUERLINGER_2007_ScalabilitySPEC.pdf

[Please note that this is the version I have originally submitted to the 
conference, i.e., it is not yet the camera-ready version that takes the 
reviewer's comments into account, etc.]

Best regards,

Karl Fuerlinger



On Thu, 8 Feb 2007, soohong kim wrote:

> Hi,
> 
> I'm looking for performance study of SPEC.OMP or any other OMP benchmarks.
> Specifically, I'm interested in seeing quantitative data of the overheads of
> OpenMP (such as barr sync, fork-join, etc).
> 
> 
> And I was told that there was a perf study presented by Sun at WOMPAT 2002
> in Fairbanks, but couldn't locate the paper on the web.  If anyone has the
> pointer to the publication, please send me the URL for that paper
> or any such performance study publication.
> 
> Thanks.
> 
> Soohong Kim
> 

-----
Karl Fuerlinger
Innovative Computing Laboratory  | Phone: +1 865 974 3052
Computer Science Department      | Fax:   +1 865 974 8296
Univ. of Tennessee, Knoxville    | Office: Claxton 320
1122 Volunteer Blvd              | www.cs.utk.edu/~karl
Knoxville, TN 37996-3450         | 35 57'17"N  83 55'42"W