Gandalf Interpreter Aborted(Core Dumped)

[speed-wagon]

Warning: parameter riemann_order was not recognised

I have tried to run many of the sample simulations provided in the tests directory via the gandalf interpreter.

running: newsim tests/feedback_tests/supernova.dat

Yields: "Warning: parameter riemann_order was not recognised" many other test simulations offer similar messages for various other parameters.

then using run to begin gandalf for the prepared parameter file causes a crash with message: "Assertion 'cell.ilast - cell.ifirst == cell.N -1' failed. Aborted (core dumped)"

I am unsure how to go about debugging this, as the error is not very descriptive and doesn't tell me much on its origin point.

Any assistance would be appreciated

[rbooth200]

The warning refers to an old parameter that is no longer used and can safely be ignored.

However, I can't reproduce your problem with the supernova test. Can you provide more information?

[speed-wagon]

Some details about my system.

• running fedora 31 on a hyper-v linux virtual environment.
• using python 2.7.17

Full message after "gandalf> run":

[SimulationBase::Setup]
[SimUnits::SetupUnits]
Finished creating initial conditions
python2.7: /home/seanbayly/Documents/gandalf/src/Tree/KDTree.cpp:462: void KDTree<ndim, ParticleType, TreeCell>::DivideTreeCell(int, int, ParticleType<ndim>*, TreeCell<ndim>&) [with int ndim = 2; ParticleType = GradhSphParticle; TreeCell = KDTreeCell]: Assertion `cell.ilast - cell.ifirst == cell.N - 1' failed.
Aborted (core dumped)

Since the error In the thread title is not the culprit I will update it.

[rbooth200]

I'm not sure whats going on. Do any of the test problems run using the binary without the interpretter?

[speed-wagon]

when running from the command line with ./src/gandalf ./tests/feedback_tests/supernova.dat

The same message is shown as in the interpreter

[rbooth200]

Can you give me more info about your build? Which branch are you on, which compiler?

[speed-wagon]

using the g++ compiler standard with Fedora 31.

[speed-wagon]

I am on the master branch

[rbooth200]

I don't know of any issues on red hat or ubuntu, it must be something peculiar to Fedora. I'm afraid I don't have Fedora 31 available at the moment, nor time to set up a virtual machine.

You could try a different compiler.

[speed-wagon]

@rbooth200 I have changed over to Ubuntu this time not on a virtual machine. most of the example simulations work fine now. that is except for the Astro sims.

For example BinaryAccretion.py run with python2.7 it manages to create the initial conditions okay. but then when it tries to begin it says :int ComputeTimeStepLevel(Double, Double): assertion 'dt > 0.0' failed

[rbooth200]

Can you confirm whether the boss-bodenheimer tests works? This is one of the most well tested.

[speed-wagon]

Yes, boss-bodenheimer does work for me

[rbooth200]

Ok, great. It seems like you have a version that is working the same as mine. I'm not sure what the issue with the binary accretion test is, I can look into it if you need it for something.