Problem reading Aspect output particles in Paraview

[UfrancoU]

Hi all,

I hope you guys are fine.

I am using ASPECT, in particular the following:

-- This is ASPECT, the Advanced Solver for Problems in Earth's ConvecTion. -- . version 2.1.0 -- . using deal.II 9.1.1 -- . with 32 bit indices and vectorization level 1 (128 bits) -- . using Trilinos 12.10.1 -- . using p4est 2.2.0 -- . running in DEBUG mode -- . running with 24 MPI processes

I am finding trouble trying to use particles when running my 3D simulations. I have successfully used particles in another project, but with an older version of Aspect, and in that case I used an 'Ascii file' as the particle generator.

In my current project, I have tried two different generators: Ascii file, and Uniform box.

The system runs, produces output. Then, when I load this output in Paraview 5.7.0, it successfully loads the solution files (for viscosity, density, Temp, etc.), but upon loading the particles.. it crashes ! If I try loading only the particles, it crashes equally. Paraview turns off !

If using Paraview 4.4, it issues an error message of this kind: duplicate indexes. And it cannot plot the particles.

I have tried prescribing many particles, few, even 1 single particle. I have also tried changing the number of grouped files. The crash is always the same.

What is the problem? What should I modify?

cheers many, Felipe

[bangerth]

The answer to these sorts of questions is always the same: Create a minimal test case that shows the problem and send us the input file that re-creates the problem.

[UfrancoU]

Hi Wolfgang, or to whom it ay concern,

I downgraded my prm file to try to catch the error on particle generation.. The error of not producing good particles persisted !

In the end I decided to download the convection_box_3D example from the Aspect hub to play with the basics. I run it, it runs well. Then I added the particles

set List of postprocessors = velocity statistics, temperature statistics, heat flux statistics, visualization, particles ...

subsection Particles set Number of particles = 16 set Time between data output = 0.0001

set Data output format = vtu

    set Particle generator name = uniform box # ascii file
    set Number of grouped files = 1
    set List of particle properties = initial position, pT path #initial composition,

    subsection Generator
        subsection Uniform box
            set Maximum x = 0.5
            set Minimum x = 0.1
            set Maximum y = 0.4
            set Minimum y = 0.2
            set Maximum z = 0.9
            set Minimum z = 0.6
        end
   end

end

It runs well, produces all the output. Upon plotting on Paraview, the particles crash, the error is ´duplicate array names´

Inspecting the runtime logfile, I realized that in my runs, with different model prm files, the number of advected particles is greater than the number I prescribed ..somehow the computer is creating more particles than the ones I prescribe. (In the example above, I prescribed 16 particles, but in the run starting from step 1 there are 18 advected particles ! ).

So, I guess there may be an error in my ASPECT-Trilinos-dealii installation. I have no idea how to solve this.

cheers, Felipe

[gassmoeller]

Hey @UfrancoU, these are two separate issues. With the default particle generator ASPECT will sometimes not generate exactly the number of particles you prescribed, because the problem of how to distribute them between different processes and cells is actually not trivial. I think this is fixed in ASPECT version 2.2 and newer, but it should now affect your output (you will simple have slightly more or less particles than you prescribed). Concerning your output crash, this may actually be a bug in the specific combination of ASPECT and deal.II you use. What happens if you remove all particle properties? Does it still crash when loading the output?

[UfrancoU]

 

 Hi Wolfgang,

 

   Thanks for replying.

 

    Following your advice on investigating, I commented (invalidated) the line for set particle properties (and Aspect 210 has no default for that).

 

    The corresponding output for solution is all good, but the particles' output makes Paraview 5.7.0 to crash, in the same way.

 

 

   Well, maybe it is as you say, a glitch on combinations of versions of dealii & Aspect.

 

   I may try to install another Aspect version if I deem it crucial.

 

    Let me know if you have other ideas on how to circumvent the issue.

 

 

thanks a lot,

cheers many,

Felipe

 

 

 

  

 

 

 

 

Sent: Friday, August 14, 2020 at 10:58 AM From: "Rene Gassmoeller" notifications@github.com To: "geodynamics/aspect" aspect@noreply.github.com Cc: "UfrancoU" aloctono@gmx.com, "Mention" mention@noreply.github.com Subject: Re: [geodynamics/aspect] Problem reading Aspect output particles in Paraview (#3685)

 

Hey @UfrancoU, these are two separate issues. With the default particle generator ASPECT will sometimes not generate exactly the number of particles you prescribed, because the problem of how to distribute them between different processes and cells is actually not trivial. I think this is fixed in ASPECT version 2.2 and newer, but it should now affect your output (you will simple have slightly more or less particles than you prescribed). Concerning your output crash, this may actually be a bug in the specific combination of ASPECT and deal.II you use. What happens if you remove all particle properties? Does it still crash when loading the output?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[bangerth]

@UfrancoU Any news on trying the newest deal.II and ASPECT versions?