I have the results from KAUST on our Lustre with 1000 compute nodes.
We did develop an “embarrassing parallel find”, basically it is a really simple C tool
where MPI process executes find on different folder. We have seen some nice results
compared to default solution, some cases from 4000 IOPs became ~61000 IOPs when we
increased significant the participated MPI processes.
George Markomanolis, PhD
KAUST Supercomputing Laboratory (KSL)
King Abdullah University of Science & Technology
Al Khawarizmi Bldg. (1) Room 0123
Kingdom of Saudi Arabia
Mob: +966 56 325 9012
Office: +966 12 808 0393<tel:%2B966%2012%20808%200683>
This message and its contents including attachments are intended solely for the original
recipient. If you are not the intended recipient or have received this message in error,
please notify me immediately and delete this message from your computer system. Any
unauthorized use or distribution is prohibited. Please consider the environment before
printing this email.