[gmx-users] How to use ambconv package...

[david.evans at ulsop.ac.uk]

Hi, Tomas,

Are you using ambconv regularly? I managed to get it working, but was
slightly concerned that the energies between AMBER and Gromacs did
not agree so well (~1% difference for isolated DNA duplex, ~10% for
VdW in a solvated DNA duplex), compared to Filip Ryjacek's stated
accuracy of 0.01%. I was wondering if you had managed to get better
accuracy.

Thanks

Dave

---- Message from Tomas Kubar <tomas.kubar at uochb.cas.cz> at
2004-06-18 12:10:02 ------
>Hi again
>
>> There is also a bug in ambconv.c -- (thanks to Anton Feenstra for
>> pointing this out).
>> - s_atop* atop; */
>>                                                                   
  
>>                                                                   
  
>>    
>> +/* DAE Insert Anton Feenstra's suggestion on memory allocation */
>>                                                                   
  
>>                                                                   
  
>>    
>> +s_atop   satop; /* static */
>> +s_atop*  atop;  /* pointer */
>>                                                                   
  
>>                                                                   
  
>>   +atop = &satop;  /* set atop to point to satop's adress */
>
>Yes, this is a possibility. Instead, my version includes
>
>s_atop* atop = (s_atop*) xmalloc(sizeof(s_atop));
>
>Futhermore, I have slightly different util.c, line 74 and following
9 
>similar lines:
>
>#anything#_filename = (char*)
malloc(strlen(argv[i+1])*sizeof(char)+1);
>
>Another problem may arise, if your system contains 1000 atoms or
more. 
>Should this be your case, I would recommend to modify the file 
>write_grotop.c to write 'longer' atom numbers when printing the
bonds, 
>angles and dihedrals, e.g. line 396
>
>fprintf(grotop, "  %4d%4d%s%12.5f%12.0f%s", ++atom1, ++atom2, "  
1", 
>bl, bfc,".");
>
>substitute with
>
>fprintf(grotop, "  %6d%6d%s%12.5f%12.0f%s", ++atom1, ++atom2, "  
1", 
>bl, bfc,".");
>
>Regards
>tomas
>_______________________________________________
>gmx-users mailing list
>gmx-users at gromacs.org
>http://www.gromacs.org/mailman/listinfo/gmx-users
>Please don't post (un)subscribe requests to the list. Use the 
>www interface or send it to gmx-users-request at gromacs.org.