Re: [AMBER] Amber20/AmberTools21 Installation Test Errors

From: Ravi Abrol <raviabrol.gmail.com>
Date: Thu, 14 Oct 2021 01:03:24 -0700

Thanks Dave.

We do plan to use the residue decomposition so I looked into the dif file.
I noticed that the file was missing and in fact nothing seems to have run
in that folder:
/usr/local/amber20///AmberTools/src/mm_pbsa/Examples/05_MMPBSA_Decomp_Residue/
Could it be that something went wrong during the tests such that the test
was not run for this case?

For the CUDA errors (not ignored), they were for either igb7 or igb8 cases.
For the CUDA parallel errors (not ignored), they were for igb7, igb8, rem,
and neb tests.

I am attaching a summary of max/relative errors for all the failed jobs
where this data is in the diff files.

Ravi

---
On Wed, Oct 13, 2021 at 6:16 AM David A Case <david.case.rutgers.edu> wrote:
> On Wed, Oct 13, 2021, Ravi Abrol wrote:
> >
> >The test diff file had multiple "possible FAILURE" lines:
> ><---
> >logs/test_at_serial/2021-10-12_00-46-31.diff
> >possible FAILURE:  check
>
> >/usr/local/amber20///AmberTools/src/mm_pbsa/Examples/05_MMPBSA_Decomp_Residue/test_statistics.out.dif
> >possible FAILURE: (ignored)  check
>
> >/usr/local/amber20/AmberTools/src/FEW/examples/test/calc_a_1t/AMT/pqr_snaps/AMT_rec.pqr.3.dif
> >possible FAILURE: (ignored)  check min.out.dif
>
> The failures marked "ignored" are expected, and you don't need to worry
> about them.  You should look at the files in the MMPBSA_Decomp_Residue
> folder to see what happened.  (But only if you are planning to use
> MMPBSA in its residue-decompostion mode.)
>
>
> The CUDA test failures are unexpected.  You should look at a few of the
> .dif
> files (and the associated outputs) to see if there is some type of problem
> that shows up repeatedly.  If you can report details of one of the CUDA
> failures, that might trigger recognition.
>
> I don't have access to a 3080Ti myself.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber

Received on Thu Oct 14 2021 - 01:30:02 PDT