[AMBER-Developers] Patches for sander.PUPIL

From: Ben Roberts <roberts.qtp.ufl.edu>
Date: Mon, 12 Jul 2010 13:33:46 -0400

Hi all,

I've attached a couple of patches for Amber 11, applicable to
src/sander/makedepend and src/sander/mdread.f respectively, which are
designed to stop the user from inadvertently running a PUPIL calculation
with implicit solvent (since that causes hideous crashes or bogus
results). Let me know if there are any problems with them.

While this is the best way I know of to submit patches in, for my own
reference, is there another, preferred way?

Cheers,
Ben

********>Bugfix 6:
Author: Benjamin P. Roberts

Date: July 12, 2010

Programs: sander (PUPIL variant)

Description: Any attempt to use implicit solvation (PB or GB) in the context
             of a QM/MM calculation with PUPIL will cause either crashes or
             nonsense results. I have added a check to mdread.f so that the
             program will stop if the user tries this combination. This change
             to sander's "makedepend" adds mdread.f to the list of files that
             differ between regular sander and sander.PUPIL.

Use this patch in src/sander/makedepend
------------------------------------------------------------------------------
--- ../../../Amber12/src/sander/makedepend 2010-05-11 15:37:14.000000000 -0400
+++ makedepend 2010-05-06 17:10:06.000000000 -0400
.. -29,7 +29,7 ..
         $LES{"evb_umb_primitive"} = $LES{"out_evb"} = $LES{"remd"} = 1;
 
 $PUPIL{"sander"} = $PUPIL{"force"} = $PUPIL{"mexit"} = $PUPIL{"mdfil"} =
- $PUPIL{"printe"} = $PUPIL{"dynlib"} = 1;
+ $PUPIL{"mdread"} =$PUPIL{"printe"} = $PUPIL{"dynlib"} = 1;
 
 $RISM{"dynlib"} = $RISM{"force"} = $RISM{"mdfil"} = $RISM{"mdread"} =
     $RISM{"new_time"} = $RISM{"printe"} = $RISM{"runmd"} = $RISM{"sander"} = 1;


********>Bugfix 7:
Author: Benjamin P. Roberts

Date: July 12, 2010

Programs: sander (PUPIL variant)

Description: Any attempt to use implicit solvation (PB or GB) in the context
             of a QM/MM calculation with PUPIL will cause either crashes or
             nonsense results. This modification to mdread.f causes the program
             to stop if the user tries this combination.

Use this patch in src/sander/mdread.f
------------------------------------------------------------------------------
--- mdread.f.bak 2010-05-06 16:09:42.000000000 -0400
+++ mdread.f 2010-05-06 16:15:29.000000000 -0400
.. -1937,6 +1937,15 ..
       inerr = 1
 #endif
    end if
+ ! BPR: PUPIL does not work with GB (or, I suppose, PB) for the
+ ! time being. It is known to either crash or produce bogus
+ ! results.
+#ifdef PUPIL_SUPPORT
+ if (igb > 0 .or. ipb /= 0) then
+ write(6,'(a)') 'Cannot use implicit solvation (GB or PB) with PUPIL'
+ inerr = 1
+ end if
+#endif /*PUPIL_SUPPORT*/
    if( (igb > 0 .or. ipb /= 0) .and. numextra > 0) then
       write(6,'(a)') 'Cannot use igb>0 with extra-point force fields'
       inerr = 1


_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Mon Jul 12 2010 - 11:00:04 PDT
Custom Search