[gmx-users] parallel problem

[Erik Lindahl]

On Mon, 2002-03-04 at 17:41, Choon Peng wrote:
> Hi All,
> 
>   I wonder if anyone tried running the benchmark suite?
> All goes well except when I tried running the Poly-CH2 benchmark for 2
> nodes:
> Fatal error: ci = -2147483648 should be in 0 .. 1727 [FILE nsgrid.c, LINE
> 210]
> Error on node 0, will try to stop all the nodes
> [0] MPI Abort by user Aborting program !
> [0] Aborting program!
> p0_6205:  p4_error: : -1
>     p4_error: latest msg from perror: Resource temporarily unavailable
> Fatal error: ci = -2147483648 should be in 0 .. 1727 [FILE nsgrid.c, LINE
> 210]
> Error on node 1, will try to stop all the nodes
> [1] MPI Abort by user Aborting program !
> [1] Aborting program!
>     p4_error: latest msg from perror: Resource temporarily unavailable
> p1_17879:  p4_error: : -1
> 
> Anyone encountered a similar error?
> Gromacs 3.0.5 did not give this error.
> All other benchmark cases work as before. (Using MPICH 1.2.3 to configure
> FFTw and GROMACS)

Hi,

No - for me it works splendid on both 1 & 2 cpus in version 3.1. Of
course it might be a bug that only shows up in mpich, but I would start
by doing a clean recompile, and perhaps see if it is present with lam
too.

Cheers,

Erik