Steve
This may happen if two jobs write to the same file at the same time. Did you have an old process from a previous run still writing output when you started a new job with the same output directory? Or is there an error in your job startup, so that both MPI processes think they are rank 0? (Look at Carpet's output; it will tell you how many ranks MPI thinks there are.)
-erik
On Thu, Jan 24, 2013 at 5:06 PM, Steven R. Brandt sbrandt@cct.lsu.eduwrote:
In some of my simulations, output with Carpet with CarpetIOASCII on 2 procs, I get the following strange output:
3 0 0 0 0 13 34 0 0.142955102930643 5 15.5 0 -0.214373883367341 3 0 0 0 0 14 34 0 0.142955102930643 5.5 15.5 0 -0.229590114556558 3 0 0 0 0 15 34 0 0.142955102930643 6 15.5 0 -0.242231416010292 3 0 0 0 0 16 34 0 0.142955102930643 6.5 47 0 0.142955102930643 18 22 0 0.18719967193547 3 0 0 0 0 40 47 0 0.142955102930643 18.5 22 0 0.198544304935133 3 0 0 0 0 41 47 0 0.142955102930643 19 22 0 0.206737466119412 3 0 0 0 0 42 47 0 0.142955102930643 19.5 22 0 0.211879729772937
and
3 0 0 0 0 31 28 0 0.142955102930643 14 12.5 0 -0.272221574961306 3 0 0 0 0 32 28 0 0.142955102930643 14.5 12.5 0 -0.271474578781231 3 0 0 0 0 33 28 0 0.142955102930643 15 12.5-0.0296722731971698 3 0 0 0 0 57 22 0 0.142955102930643 27 9.5 0 -0.0249091567790721 3 0 0 0 0 58 22 0 0.142955102930643 27.5 9.5 0 -0.0207705868267214
(Note the missing delimiter between 12.5 and -0.0296722731971698)
From the par file:
IOASCII::out2D_every = 1 IOASCII::out2D_xyplane_z = 0 IOASCII::out2D_vars = "FunWave::eta FunWave::u FunWave::v" IOASCII::out2D_xz = "no" IOASCII::out2D_yz = "no" IOASCII::output_ghost_points = "no"
Does anyone know what this problem is?
Cheers, Steve
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users