Muskingum: mRM initialization takes much more time in release 5.9 than in release 5.8 (measurable only for large domains)

Dear all,

Given some inputs from an external user and also confirmed by Pallav, there seems to be a non-negligible time difference between mRM initialization for routing case 1 (=Muskingum) between release 5.8 and 5.9.

I have checked recently myself on one EDgE basin (~120000km2), 30years of simulations, compiled with intel18.

mHM release 5.8: total runtime 13minutes mHM release 5.9: total runtime 22minutes (~11minutes MPR initialization)

Same I did for a smaller basin (~25000km2), the differences were smaller, but also appeared.

Does someone of you have a clue?

Lastly, these are no differences in runtime for the adaptive time stepping (routing case #2 (closed))

Thanks!

Olda