Fabian Fröhlich

This definitely looks like a bug, thanks for such a detailed report!

Hi Matthias, I am still not sure whether I completely understand the issue you are describing, but I am no longer convinced the specific setup you are describing here is...

Thanks, but the primary issue is that we need different behaviour depending on the format and entries of the condition table, i.e., depending on whether only compartment size or also...

Well technically even if the process runs on a single core it could still use multiple threads via hyper-threading, but overall this is interesting and I wasn't aware of it....

fixed in https://github.com/AMICI-dev/AMICI/commit/ed36b2daa214029487588a50b77447d3dac76f78

unrelated, as in "we changed the way the solver times tested in these tests are calculated in https://github.com/AMICI-dev/AMICI/commit/f90916605750e58021c3481a546f09256c9dbf8e"?

> > unrelated, as in "we changed the way the solver times tested in these tests are calculated in [f909166](https://github.com/AMICI-dev/AMICI/commit/f90916605750e58021c3481a546f09256c9dbf8e)"? > > This is a different time, not used anywhere...

> Would be nice to also show (and possibly check) numsteps, nrhsevals, ... there. agreed

Agreed, should be part of condition table.

👍