Hello,
after writing my bachelor-thesis on "Benchmarking non-blocking I/O
operations on compute clusters" I am due to give a presentation on
"Benchmarking Parallel File Systems" in general.
In order to give the presentation a little up-to-dateness and closeness
to reality, I was wondering if there are benchmarking experiences
(problems, ideas on what works great/better/not at all...) around which
you might share with me.
What kind of benchmarks do you as PVFS2-Users (or potential users) use
to decide if this is the FS of your choice? Recently the question on
IOzone and PVFS2 arose, what else do you use?
How do you compare PVFS2-performance to the performance of other
available Parallel File Systems?
How well do you think do available Benchmarks represent your usual
access patterns? Do you use those benchmarks or do you create benchmarks
"in house" for your own special purpose?
It would be awesome if anyone with any kind of experience (good as well
as bad :-) ) in this domain with a little time could give me a bit of
information about this. I think it would be great to be able to use this
information to underline or to disprove theory.
Thanks a lot,
Regards David
_______________________________________________
Pvfs2-users mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users
