On Fri, Aug 10, 2012 at 8:35 AM, case <case.biomaps.rutgers.edu> wrote:
> On Wed, Aug 08, 2012, Raik Grünberg wrote:
>> >
>> > Currently, the sqm program only prints Mulliken charges to
>> > three digits after the decimal place, which can lead to
>> > normalization errors. You could try to modify sqm to print out
>> > more digits and see if that helps. (Look around line 47 in
>> > $AMBERHOME/AmberTools/src/sqm/qm2_print_charges.F90.
>>
>> Thanks for the hint. I have never programmed in Fortran so i just made
>> a guess and changed
>> line 43 from:
>> write(6,'(" ",i5," ",A2," ",F14.3," ",F14.3)') i, &
>> to:
>> write(6,'(" ",i5," ",A2," ",F14.6," ",F14.6)') i, &
>>
>> and applied the same change to line 47 so that it now reads:
>> write(6,'(" ",i5," ",A2," ",F14.6)') i, &
>>
>> ... and recompiled everything and re-run antechamber.
>>
>> Unfortunately, that didn't make any difference. The new mol2 file is
>> identical to the old one.
>
> Is the sqm.out file changed? That is, are there now 6 digits in the Mulliken
> charges? I have only limited computer access right now, so I can't do more
> checking myself.
My workstation has been hospitalized. The last Ubuntu update seems to
have trashed the NVidia driver and the usual recompiling doesn't work.
I hope I can check on Monday. What I do still remember though is that
the overall Mulliken charge reported in the sqm.out was always
-1.00000 to at least 6 digits. I thought the partial charges had also
already been 6 digits even without the modification of the code. But I
am not 100% sure about that.
Greetings
Raik
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
___________________________________
Raik Grünberg
http://www.raiks.de/contact.html
___________________________________
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 10 2012 - 20:30:04 PDT
