search

alekseyzimin / masurca

GNU General Public License v3.0
243 stars 35 forks source link

IOPS requirements for MaSurca run #177

Closed mahalel closed 3 years ago

mahalel commented 4 years ago

Hi I am trying to look at best/cheapest options for underlying storage for MaSurca jobs.

Analyzing a run that is going on at the moment (in mr_pass2 stage) - I seem to conclude that high disk IOPS is not necessary as the data seems to be loaded in memory and only written to file once once the jf_aligner job is completed.

Which makes me wonder, is a similar process for other stages too? Or are there any stages where a file is kept open and multiple writes committed to it?

I can't find much detail on this in the documentation, would appreciate some clarity if possible please.

Regards, Andrei.

alekseyzimin commented 3 years ago

The stage in mr_pass2 is unique. MaSuRCA is not very disk intensive overall.