Hello everyone,
In light of the smooth, successful release of AmberTools 1.3, I
thought it would be appropriate to jump right into some amber11 test
issues... The biggest issue that I'm running into right now is either
a change in a default value (of ntx and/or irest), or a new test for
conflicting values of those two variables. Many tests use ntx=7 (for
which the documentation is sadly insufficient, as the manual provides
no description to tell me how it's different from 5 or 6), and sets no
value for irest (which defaults to 0). This now gives (NTX / IREST
inconsistency *** input errors).
The Run.* files "catch" these errors and jump over to an error
message, printing an uninformative "Program Error" and ending the test
(thereby making a complete test of AMBER a nightmare). Because this
occurs in many locations, it is becoming time consuming (and
mind-numbing) to essentially run the test, find where the error
occurred, verify this is the cause, comment out that specific test,
and resume. This has happened on LES_CUT, LES_GB, and more than half
of the test.sander.PIMD.partial tests.
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. Any thoughts?
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 Mon Dec 21 2009 - 21:30:02 PST
