| WMDE-leszek added a comment. |
Thanks @Smalyshev for looking into this! I've just submitted a new patchset of https://gerrit.wikimedia.org/r/#/c/345332 where I've introduced a cache to reduce number of calls to EntityId::getLocalPart and EntityId::getRepositoryName.
As I am really not an expert with the matter, if you had any suggestion what would be a good default value for the cache size (i.e. how many IDs should be cached), that would be appreciated. I guess a thing to consider would be to have a mildly reasonable "general" default (say 1K, 10K, etc), and set it to a bigger number in settings for wikidata(s) so e.g. properties would most likely be all cached.
With some sane default we could run the benchmark again and see how this helped. If you have any other suggestion, also please say, and I'll try to address them.
Also, if this makes any sense, the code could possibly also changed so that properties are cached separately, so we're sure each property ID is always processed only a single time.
Cc: hoo, Ladsgroup, PokestarFan, Lucas_Werkmeister_WMDE, Smalyshev, daniel, WMDE-leszek, Aklapper, GoranSMilovanovic, QZanden, Izno, Wikidata-bugs, aude, Mbch331
_______________________________________________ Wikidata-bugs mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
