Hi Ricardo, On Fri, 12 Mar 2021 at 12:23, Ricardo Wurmus wrote: > So it’s not unsurprising that it takes longer than “guix environment > --help”, which doesn’t have to talk to the daemon at all. The question > is just if we can avoid some of that work and if so how much of it. Thanks for all the explanations. My point is the cold vs hot cache changes a bit the analysis. --8<---------------cut here---------------start------------->8--- $ sudo sh -c 'echo 3 > /proc/sys/vm/drop_caches' $ time guix environment --help >/dev/null real 0m1.148s user 0m0.804s sys 0m0.125s $ time guix environment --help >/dev/null real 0m0.534s user 0m0.649s sys 0m0.044s --8<---------------cut here---------------end--------------->8--- --8<---------------cut here---------------start------------->8--- $ sudo sh -c 'echo 3 > /proc/sys/vm/drop_caches' $ time guix environment --profile=$HOME/.guix-profile -- exit real 0m2.087s user 0m1.681s sys 0m0.151s $ time guix environment --profile=$HOME/.guix-profile -- exit real 0m1.250s user 0m1.445s sys 0m0.126s --8<---------------cut here---------------end--------------->8--- In both cases, half the time is for “doing nothing“ and this “doing nothing” is currently the major part of the slowness, especially on cold cache: compare the attached graph. Cheers, simon