[AMBER-Developers] AmberTools19: still a (short) to-do list

From: David A Case <david.case.rutgers.edu>
Date: Fri, 26 Apr 2019 15:41:18 -0400

Hi everyone:

I've gone ahead and updated the web page to reflect AmberTools19. The
History page also shows the group photo from last month.

Thanks to everyone who pitched in a lot of work to test, debug release
scripts, and otherwise help out in getting this release out the door.

Things are not perfect: that's why we have an update mechanism. Two
problems seem generic, and not specific to a particular compiler:

  a. The "rem_wat_pv" and "hrem_wat_pv" tests are failing in at -np 4
     with pmemd.MPI. I'm not sure if these tests are supposed to work
     with pememd.MPI from Amber18 or not. If they should work, we need
     to debug. If these are testing features not due in pmemd until
     next year, we need to comment out the tests in the released version.

  b. The "cnstph_remd/Explicit_PHrem" test is failing with sander.MPI
     on 4 threads. The test seems to be running twice. People involved
     in constant pH remd code should look at this.

There are some other glitches (some roundoff errors, some other things)
that are worth fixing if someone figures out what is going on. Going
forward, we need to make a cleaner separation between Amber (i.e. pmemd)
and AmberTools. But the next non-joint release is not until 2021, so we
have plenty of time to think about this.

The master branch is again "open" to start proposing merge requests.
Please don't figure you can just wait until a year from now to do
anything: the sooner code gets into the master branch, the more it gets

The amber19-with-patches branch should correspond very closely
(hopefully identically except for things like devtools that we don't
distribute) to what users have. Follow the bugfix instructions on the
wiki to get updates out to users.

...thanks again to all....dac

AMBER-Developers mailing list
Received on Fri Apr 26 2019 - 13:00:02 PDT
Custom Search