the reason is that by default poetry puts virtual environments in its cache directory
so while all the other tools are working in an existing virtual environment but with empty cache, poetry is working from a completely fresh start
There's definitely a case to be made that this is poetry's own silly fault, if it didn't want people who clear the cache also to clear their venv then it shouldn't have done that.
But I think it is more in the spirit of what the benchmark is trying to measure to keep these things separate, and keep the tests more comparable.
per https://github.com/lincolnloop/python-package-manager-shootout/pull/15#issuecomment-1950141763
There's definitely a case to be made that this is poetry's own silly fault, if it didn't want people who clear the cache also to clear their venv then it shouldn't have done that.
But I think it is more in the spirit of what the benchmark is trying to measure to keep these things separate, and keep the tests more comparable.