I want to manage a clustered computation on raster and thus use the "clusterR" function in R (package 'raster'). However, there's a tuning m parameter to set, but the only explanation about this parameter is

m : tuning parameter to determine how many blocks should be used. The number is rounded and multiplied with the number of nodes.

But which value must be associated to this parameter ? How to decide it ?

  • You can use system.time() function to assess performance with different values of m to find out which value is more appropriate for your system. Nov 7 '14 at 9:52
  • ok but i'm using this clustering in many different situation and I don't "want" to define by trials&error the best value... any idea to find the "good parameter" knowing the number of core and for example the size of raster to deal with ?
    – nmatton
    Nov 27 '14 at 11:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.