On 11/17/2014 07:56 PM, Zac Medico wrote: > For hitomi, _slot_operator_update_probe/use_reduce is an obvious thing > to optimize. It called use_reduce 53763 times there, so it seems to > repeat use_reduce multiple times for the same packages. That means we > should see a benefit from memoization.
I've written a patch for this [1], and it gives good results (22.4% less time for dep calc on one of my computers). If you do more profiling, it would be best to apply this patch first, in order to increase the contrast for other hot spots. [1] https://bugs.gentoo.org/show_bug.cgi?id=529660 -- Thanks, Zac
