Hi Muhammad,
The test case you mention, tgtmd/change_target.ntr, has only 3
residues (ALA, ALA, TYR, at least according to the Amber9 test case,
which I assume is the same), which is less than the number of
processors (4). The parallel version of sander works by dividing
residues among processors, which needs more residues than processors
to work. So, the error mesage only means that sander cannot divide 3
residues in 4 processors, and then it stops. This test should work
fine with 2 processors though.
Gustavo.
On 1/23/07, Muhammad Naim Mohmad Rouyan <m_naim.putra.upm.edu.my> wrote:
> Dear All,
>
> I have compiled AMBER8 and run test.sander and found errors.
>
> When I run using 'mprun -np 2' the result is
>
> ==========================================================
> cd tgtmd/change_target.ntr; ./Run.tgtmd
> SANDER: Targeted MD with changing target and restraints
> diffing tgtmd.out.save with tgtmd.out
> PASSED
> ==========================================================
>
> Then when I run using 'mprun -np 4' the result is
>
> ==========================================================
> cd tgtmd/change_target.ntr; ./Run.tgtmd
> SANDER: Targeted MD with changing target and restraints
> Fatal error, aborting.
> Must have more residues than processors!
> Must have more residues than processors!
> Fatal error, aborting.
> Must have more residues than processors!
> Fatal error, aborting.
> Fatal error, aborting.
> Job cre.735 on inspem1: received signal ABRT.
> ./Run.tgtmd: Program error
> *** Error code 1
> make: Fatal error: Command failed for target `test.sander.no_lmod'
> ==========================================================
>
> I don't know why this error occur. Does this test no support parallel
> greater then 2.
>
> Regards....naim
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Jan 24 2007 - 06:07:40 PST
