[gmx-users] with PME pot. energies = nan !! - Is that FFTW ??

[Arturas]

Hi,

If I prepare em.mdp (energy minimisation,steep) with 'coulombtype =
PME', than each step of mdrun produces 'Potential energy = nan'. And
other energy characteristics like kinetic , atc. also are '=  nan'. If
I use other coulombtype - energies is ok'.

In manual i noticed that PME uses FFTW, so decied could it be related
with too old version of FFTW to GROMACS 3.2.1.




|----->
DvdS> On Fri, 2004-09-24 at 12:25, Arturas wrote:
>> Hello,
>> 
>> I have managed to find that using PME for electrostatics energies are
>> not produced, e.g. making energy minimisation potential energy = nan.
>> Other 'coulombtype' options works fine.
DvdS> can you be more specific...

>> 
>> Can this be related to old FFTW version if GROMACS was reinstalled ?
>> Is there anything more about that ?
>> 
<-----|