AMBER Archive (2009)Subject: [AMBER] make test.parallel.MM error
From: Collins Nganou (nganoucollins_at_gmail.com)
Date: Sun May 03 2009 - 16:20:51 CDT
Thanks David,
After set export DO_PARALLEL='mpirun -np 4'
I receive the same error as before. I think the variable DO_PARALLEL is not
recognised by amber10.
make[1]: Leaving directory `/opt/amber10/test'
cd rem_gb_2rep && ./Run.rem
REM requires a parallel environment; skipping this test
cd rem_gb_4rep && ./Run.rem
REM requires a parallel environment; skipping this test
cd rem_wat && ./Run.rem
REM requires a parallel environment; skipping this test
cd rem_hybrid && ./Run.rem
REM requires a parallel environment; skipping this test
cd sodium && ./Run.sodium
TI can only be run parallel: set env var DO_PARALLEL
Not running test, exiting.....
cd pheMTI && ./Run.0
diffing out.0.save with out.0
PASSED
==============================================================
cd pheMTI && ./Run.1
diffing out.1.save with out.1
PASSED
==============================================================
cd pheMTI && ./Run.lambda0
This test must be run in parallel
make: *** [test.sander.TI] Error 1
yours faithfully
Collins
On Thu, Apr 30, 2009 at 10:16 PM, David A. Case <case_at_biomaps.rutgers.edu>wrote:
> On Thu, Apr 30, 2009, Collins Nganou wrote:
>
> > After doing some investigation, I discover that the variable
> DO_PARALLEL
> > is not recognised by Amber10.
> > The variable was properly declared in my bashrc file as follow : alias
> > DO_PARALLEL='mpirun -np 4'
>
> This is *not* the proper way. You need to set the DO_PARALLEL environment
> variable, not a shell alias. For bash, type:
>
> export DO_PARALLEL='mpirun -np 4'
>
> (N.B.: this variable is only used for the test cases. When you run your
> own
> jobs, you need to call mpirun in the ordinary way.)
>
> ...dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
|