[Beowulf] file IO benchmark

[Joachim Worringen]

Toon Knapen wrote:
> The problem is that our parallel direct out-of-core solver thus needs to
> store tmp data on disk. We already encountered problems when people are
> using one global NFS mounted filesystem for storing the tmp data of all
> nodes in the parallel run. Of course this results in continuous
> conflicts on the (slow) ethernet network and a disk trying to stream
> multiple big files at the same time. Requesting users to run this
> 'standard' benchmark will enable us also to detect these kind of situations.

I like b_eff_io (based on MPI-IO) as it it easy to run but nevertheless produces 
a sufficently large range of number for different parallel(!) access stlyes and 
patterns like accessing independent or shared files etc. Also, it is quite smart 
as it allows to limit the execution time and produces quite reliable numbers. 
Even if you don't use MPI-IO in your application (you should...), the numbers 
for non-collective file access are valuable.

Google leads your way to download.

  Joachim

-- 
Joachim Worringen - NEC C&C research lab St.Augustin
fon +49-2241-9252.20 - fax .99 - http://www.ccrl-nece.de