We maintain some conda environments using the intel channel. Working with our python 2.7 environment is incredibly painful. We will see:
# conda update -n scipy -c intel python=2.7 --all Solving environment: \
take 30-290 minutes on fast hardware. I'm assuming this may be due to a large number of available packages or some particular conflicts. Is there any chance that this can be reduced?
Can you send me the output of "conda info" and "conda list --explicit" from an environment exhibiting the problem? I will attempt to reproduce the slow resolution.
Some simple things to try yourself are clearing all local caches (conda clean --all) and paring your channel list to the minimum. Also, if there are key packages you need with specific versions, you could pin those before running the update, thereby pruning the solver's search space. Don't leave them pinned unless it makes sense for your general use case.
I've managed to speed things up quite a bit by limiting the channels with:
conda update -n scipy --override-channels --channel intel --channel main --channel anaconda-nb-extensions --all
So the issue was likely with some other default channel. Thanks for the response.