In the last part ("the real real life"), you say 30% of the time is take by C4::Context::preference, but I don't see where this number comes from. If I look at the flame graph of the "second profile" (https://joubu.github.io/koha/nytprof/dbic/nytprofhtml.21326_2/) C4::Context::preference only takes 2.42% In the last profile (which is with your patch, right ?), C4::Context::preference exclusive's time is bigger than any other subroutines and we spend 9.22s (inclusive) inside C4::Context::preference for "only" 283 calls, which seems huge. So... maybe I don't understand correctly NYTProf output, or there is something wrong with your patch Le 11/01/2016 15:57, Jonathan Druart a écrit :
Hi devs,
I have played with NYTProf at the end of the last week, here is my loosely structured notes: https://joubu.github.io/koha/nytprof/dbic/index.html
I have not found big things to implement in order to improve the speed but I am wondering if there is not something wrong with the queries cache (Search in the page for "Update:")
Hope to get some feedbacks.
Cheers, Jonathan
PS: If someone has a better idea to share the profiles, I am open-minded. _______________________________________________ Koha-devel mailing list Koha-devel@lists.koha-community.org http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel website : http://www.koha-community.org/ git : http://git.koha-community.org/ bugs : http://bugs.koha-community.org/
-- Julian Maurice <julian.maurice@biblibre.com> BibLibre