[Users] interpolation error

[林家暉]

Dear Rolnd and Zach,

Thanks for your explanation. My domain size is (-540,-540,-540)x(540,540,540), so the interpolation point is outside of the domain as you said.

For using out_of_bounds_xyz, do I need to rerun the simulation from the beginning or I can run it from the last checkpoint after adding out_of_bounds_xyz in the parameter file ?

Best regards,

Chia-Hui

________________________________
寄件者: Zach Etienne <zachetie at gmail.com>
寄件日期: 2019年5月30日 下午 11:03:03
收件者: Haas, Roland
副本: 林家暉; Einstein Toolkit Users
主旨: Re: [Users] interpolation error

Hi Chia-Hui,

particle_tracerET already has a parameter for this: out_of_bounds_xyz

I follow the general rule of setting it to about 80% of my (coarsest) grid size.

-Zach

*     *     *
Prof. Zachariah Etienne
West Virginia University
https://math.wvu.edu/~zetienne/
https://blackholesathome.net


On Thu, May 30, 2019 at 10:58 AM Haas, Roland <rhaas at illinois.edu<mailto:rhaas at illinois.edu>> wrote:
Hello Chia-Hui,

the error message indicates that the code tried to interpolation a
quantity to location (x,y,z)=(-157.725,-576.349,-99.2027) but that the
grid only has a size of (xmin,ymin,zmin)x(xmax,ymax,zmax) =
(-576,-576,-576)x(576,576,576) (where I made a guess for the ymax and
zmax values given that the output only shows the shape of the grid on
one MPI rank).

So the interpolation point is outside of the domain.

Depending on what caused the interpolation this could mean eg that a
tracer particle (given the function name
"DoOneRK4StepForParticleTracerET") tried to leave the domain. This is
something that needsto be fixed in the particle_tracerET thorn,
probably by making it deactivate particles that get too close to the
edge of the domain.

Yours,
Roland

> Hello,
>
> I am using IllinoisGRMHD with the test with nsns_test.par in https://bitbucket.org/zach_etienne/wvuthorns_diagnostics/src/master/NSNS_parameter_files/  .
>
> The simulation runs well until t~5300 and an error occurs. The error message is like this:
>
>
> WARNING level 1 from host nid02210 process 0
>   while executing schedule bin particle_tracerET, routine particle_tracerET::DoOneRK4StepForParticleTracerET
>   in thorn AEILocalInterp, file /global/u1/c/chlin/et/Cactus/arrangements/Numerical/AEILocalInterp/src/Lagrange-tensor-product/../template.c:1109:
>   ->
>    CCTK_InterpLocalUniform():
>         interpolation point is either outside the grid,
>         or inside but too close to the grid boundary!
>         (this may be caused by a global interpolation with
>          driver::ghost_size too small)
>         0-origin interpolation point number pt=0 of N_interp_points=1
>         interpolation point (x,y,z)=(-157.725,-576.349,-99.2027)
>         grid x_min(delta_x)x_max = -576(18)576
>         grid y_min(delta_y)y_max = -576(18)54
>         grid z_min(delta_z)z_max = -576(18)54
>
> WARNING level 1 from host nid02210 process 0
>   while executing schedule bin particle_tracerET, routine particle_tracerET::DoOneRK4StepForParticleTracerET
>   in thorn particle_tracerET, file /global/u1/c/chlin/et/Cactus/arrangements/WVUThorns_Diagnostics/particle_tracerET/src/Interpolate_velocities_at_particle_positions.C:131:
>   -> interpolation screwed up
> WARNING level 1 from host nid02210 process 1
>   while executing schedule bin particle_tracerET, routine particle_tracerET::DoOneRK4StepForParticleTracerET
>   in thorn particle_tracerET, file /global/u1/c/chlin/et/Cactus/arrangements/WVUThorns_Diagnostics/particle_tracerET/src/Interpolate_velocities_at_particle_positions.C:131:
>   -> interpolation screwed up
> srun: error: nid02211: tasks 2-3: Exited with exit code 1
> srun: Terminating job step 21753671.0
> srun: error: nid02210: tasks 0-1: Exited with exit code 1
>
>
> I think maybe t can be avoided by decreasing the timestep so that the tracer particle would not move too fast to make the  interpolation screwed up but I am not sure. So what I want to ask is :
>
>   1.  It this error can be avoided by decreasing the timestep ?
>   2.  If so, how can I  decrease the timestep in parameter file (or in other places)?
>
>
> Thank you.
> Best regards,
> Chia-Hui



--
My email is as private as my paper mail. I therefore support encrypting
and signing email messages. Get my PGP key from http://pgp.mit.edu .
_______________________________________________
Users mailing list
Users at einsteintoolkit.org<mailto:Users at einsteintoolkit.org>
http://lists.einsteintoolkit.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.einsteintoolkit.org/pipermail/users/attachments/20190530/dea36114/attachment-0001.html