We tend to use IOR from http://sourceforge.net/projects/ior-sio/
Some tests, how we used it, and data available at http://www.cs.sandia.gov/rsiopa/ --Lee On Mon, 2008-02-11 at 09:30 -0700, Jeremy Mann wrote: > On Mon, 2008-02-11 at 11:21 -0500, mayur bhosle wrote: > > hi everyone, > > > > I am a student at Georgia Tech university, and > > as a part of a project i need to benchmark lustre file system. I did a > > lot of searching regarding > > the possible benchmark, but i need some advice on which benchmarks > > would be more suitable... if any one can post a sugesstion that would > > be really helpful....................... > > In our environment, I used iozone to benchmark our bonded gigE OSTs. > > iozone -t1 -i0 -il -r4m -s2g > > You can change -s2g to -s1g or -s512m to benchmark reads and writes to > different sizes of files. We tuned our system for large files since we > place the NCBI databases in /lustre and access them directly > within /lustre. > > -- > Jeremy Mann > [EMAIL PROTECTED] > > University of Texas Health Science Center > Bioinformatics Core Facility > http://www.bioinformatics.uthscsa.edu > Phone: 210-567-2672 > > _______________________________________________ > Lustre-discuss mailing list > [email protected] > http://lists.lustre.org/mailman/listinfo/lustre-discuss > _______________________________________________ Lustre-discuss mailing list [email protected] http://lists.lustre.org/mailman/listinfo/lustre-discuss
