I'm not quite sure where to raise this issue- I have an mf6 model that was originally developed using mf6 version 6.1.1. It is steady state and has 1 stress period (10950 days) and 6 time steps (1825 days each). With Mf6 version 6.1.1 it takes around 16 seconds to run, whereas with version 6.3 it now takes 33 seconds to run. There is no difference between the input files or the laptop I'm using to run the model (on the C: drive).

Just wondering where the overhead comes from in v 6.3? The extra time isn't a big deal (apart from using it in PEST)- it was originally a mf-nwt model converted to mf6 (using the converter executable). I converted it because it originally ran much faster in Mf6 (v 6.1.1). I now need to use the GWT functionality, hence the update to v 6.3.

Thanks for your help.

Any plans to use conda, or another packaging tool to distribute and install MODFLOW binaries? I would be willing to give this a shot to make the binaries available through conda-forge.

The version of MT3D-USGS in the win 64 download is 1.0 instead of 1.1. I have not checked the version in other downloads so it is possible that the same issue exists with them.

Is the libmf6.so file necessary to distribute along with the mf6 binary now? I don't recall this in previous releases. I am able to run ./mf6 --version without it but want to know if I need to copy it around for other purposes?

Thanks
Mike