On Tue, Dec 22, 2009 at 8:53 AM, David A. Case <case.biomaps.rutgers.edu> wrote:
> On Tue, Dec 22, 2009, Jason Swails wrote:
>
>> Many tests use ntx=7
>
> This should be fixed, although it is equivalent to ntx=5.
>
Good to know. Should I update tests that have ntx = 7 to have ntx = 5
instead? (I'm guessing this is isolated to sander/pmemd tests, though
it may affect NAB tests as well?)
>> This now gives (NTX / IREST inconsistency *** input errors).
>
> I've updated mdread.f so that the inconsistency error appears (to flag errors
> we need to fix) but does not stop the calculation (so that one can more easily
> work to correct problems.) Sorry for the problems.
Great -- once we find all of the problems in the tests (which should
pop up in TEST_FAILURES.diff due to the new flag you inserted), do you
want me to change to ntx = 5 and add irest = 1 and rerun (committing
the new output and input file(s))?
>>
>> I quickly added irest=1 to the md.in file in the LES_CUT directory and
>> the program ran fine, but got results significantly different than
>> those in the saved output file.
>
> No surprise. The reason for the change is that the code does not correctly
> use the velocities that are read in if irest=0.
Then after all of the tests are fixed, we should probably make it quit
the calculation again? There's no point in running a calculation we
know off the bat is going to give us erroneous results.
Comments?
Thanks!
Jason
--
---------------------------------------
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
352-392-4032
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Dec 22 2009 - 09:30:02 PST
