[gmx-users] pme_order and fourierspacing

[Christoph Freudenberger]

hi there,

It has been noted in the manual as well as here in the list, that
the scaling of pme in parallel execution can be improved by
increasing pme_order and fourierspacing.

I'm just playing around with the values a little bit, but I'm
not quite sure which is the best option.
I just tried

pme_order = 8 / fourierspacing = 0.48

vs

pme_order =10 / fourierspacing = 1.20

but i doesn't see to make much of a difference.

Is this dependent on the number of atoms or the system size? (my test system
contains 14000 atoms in (10nm)^3)

Which values do you usally choose?

-- 
Christoph Freudenberger
Abt. Organische Chemie I AK Siehl - Uni Ulm -Tel: ++49-731-502-2785