Hi,
I have recently installed 0.13.2 but ran into problems with more complex geometries. Simpler geometry descriptions still seem to be OK. You can see the error message below - the equivalent run in 0.13.1 on the same architecture proceeds to xs processing and runs fine.
In running a different, also complex geometry, I face similar problems. Again, this is is fine in 0.13.1. Error reported in this case: ‘*** Error in `openmc’: corrupted size vs. prev_size: 0x00000000057c7f30 ***’
Has anyone else faced similar issues? Unfortunately cannot share this exact geometry but should be able to create something if needs be.
Thanks,
Alex
Hi @alexvalentine. We did implement some geometry optimizations in version 0.13.2 for complex geometries, but it looks like something may have gone awry for your particular case. If you’re able to share a model on which this can be reproduced, we can try to get to the bottom of it.
Hi Paul, sorry for the delay - I haven’t had chance to reproduce the error with another geometry however one thing I have found is that the same geometry converted with openmc conversion tool (GitHub - openmc-dev/openmc_mcnp_adapter: Tool for converting MCNP input files to OpenMC classes/XML) actually runs OK. The original models were converted with csg2csg… Tomorrow I should get chance to see what difference in the output models is the origin of this.
Alex
Interesting! Let me know if you’re able to narrow it down to a particular cell definition. Even if there’s an error in the definition, ideally OpenMC should abort a little more gracefully than what you observed, so we may have some bug fixing to do.