Did the REMD test case pass?
you really need to give more information, it's impossible to help without
knowing more about what you are doing. let us know if the test case worked
or not and we can try to fix it from there.
On 2/23/07, Seongeun Yang <seongeun.korea.ac.kr> wrote:
>
> Hello all,
>
> After I failed to run REMD using 32 replicas with amber8 as posted a few
> days ago,
> I tried the same REMD job with amber9.
>
> The parallel version of amber9 was installed without error on Intel Xeon
> cluster,
> and the MPI version of sander with no replica exchange did the job without
> problem.
>
> But a number of attempts to run the REMD job gave the error messages such
> as below.
>
> .....
> 0 - MPI_COMM_RANK : Null communicator
> [0] Aborting program !
> [0] Aborting program!
> 1 - MPI_COMM_RANK : Null communicator
> [1] Aborting program !
> [1] Aborting program!
> p1_16499: p4_error: : 197
> p3_16533: p4_error: : 197
> 3 - MPI_COMM_RANK : Null communicator
> [3] Aborting program !
> [3] Aborting program!
> p21_1181: p4_error: : 197
> 19 - MPI_COMM_RANK : Null communicator
> [19] Aborting program !
> [19] Aborting program!
> p19_3303: p4_error: : 197
> p0_16494: p4_error: : 197
> rm_l_1_16512: (1.417969) net_send: could not write to fd=5, errno = 32
> rm_l_3_16547: (1.359375) net_send: could not write to fd=5, errno = 32
> .....
> .....
>
> Please let me know how to fix this problem.
>
> Thanks a lot.
>
> Seongeun
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Feb 25 2007 - 06:07:36 PST
