On Wed, Mar 20, 2019, Mariana Simões Ferreira wrote:
>I've been trying to install AmberTools18 following the steps in the manual
>(section 2, steps 1-7, selecting as compiler gnu) in my Ubuntu 18.04.2 LTS,
>all up to date packages. However, I've been getting failures and errors
>when I run make test (at serial, later I want to install the parallel
>option as well). Attached there are the test log and diff files.
>
>2204 file comparisons passed
>2 file comparisons failed
>6 tests experienced errors
>
>>From the Test log file, failures are the following:
>
>1) cd qmmm_EXTERN/pure_QM_MD_NWChem && ./Run.h2o.blyp_sto-3g
>diffing h2o.blyp_sto-3g.out.save with h2o.blyp_sto-3g.out
>possible FAILURE: check h2o.blyp_sto-3g.out.dif
>==============================================================
>2) cd qmmm_EXTERN/pure_QM_MD_NWChem && ./Run.h2o.blyp_sto-3g_md
>diffing h2o.blyp_sto-3g_md.out.save with h2o.blyp_sto-3g_md.out
>possible FAILURE: check h2o.blyp_sto-3g_md.out.dif
The two "failures" have to do with NWChem external QM/MM calculations.
These tests are rarely run, since few users will have NWCHem installed.
If you plan to use this option, please check the .dif files mentioned
above.
Note that some places where the word "error" appears in the log are not
errors at all: they are tests where the code is supposed to recognize
problems. For real errors, you need to look for the string "program
error" in the logs.
>2)
>if [ -n "testrism" ]; then cd ./rism1d/spc-kh && ./Run.spc; fi
>ERROR> could not open ../../../../dat/rism1d/model/SPC.mdl: 2
>STOP 1
You are getting many of these where the SPC.mdl file is not present
in AMBERHOME/dat/rism1d/model. It is actually present in the
AmberTools18.tar.bz2 file, but was removed in update.6. It's not clear
to me right now why some of the tests are still refering to these files.
I hope that people on the list (like Tyler Luchko, cc-ed here) will
recognize this. If you look at the prepdat.pl file in that folder, it
indicates that the mdl files have been deleted in 2018, but I'm unsure
right now about what is happening with the tests.
Be sure that you have run the update_amber script in your AMBERHOME
directory, if you have not done so. (cd $AMBERHOME; ./update_amber
--help for instructions.
>
>4) cd qmmm_EXTERN/QMMM_MD_NWChem && ./Run.aladip.hf_sto-3g
> Error no. 1 in getmem memory overflow : call no.,
>amount requested : 84 75250
> ./Run.aladip.hf_sto-3g: Program error
This is another NWCHem failure.
Bottom line: you should study the NWChem outputs if you plan to use that
functionality. The errors may be benign or worrisome.
It is very unlikely that you actually have a problem with rism1d. Some
path has led to an inconsistency in your tree that I don't recognize
right now. But you can continue on, using the codes and building the
parallel versions, etc.
...regards...dave case
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Mar 20 2019 - 13:30:02 PDT
