Fork me on GitHub
#clojure-nl
<
2019-02-26
>
helios07:02:00

howdy there

mpenet08:02:36

Morning. Yak shaving, trying to move from ido to ivy/counsel/swiper

borkdude09:02:36

Iā€™m playing with an optimization for clojure.core/merge. In maps with lots of elements it can lead to a speedup of 1.23x

borkdude09:02:30

with this is starting to pay off with input maps > 64 elements or so

helios09:02:04

do share šŸ˜„

borkdude09:02:37

see #clojure-dev channel

šŸ‘ 10
borkdude09:02:40

so for a 2000 element map it would mean a speedup of 35%

borkdude09:02:53

but I wonder if in practice that happens a lot šŸ˜‰

erwinrooijakkers12:02:52

From December 1st to Christmas it happens a lot ;)

erwinrooijakkers14:02:56

Hmm actually I haven't used a 2000 element map yet, even for aoc

skuro10:02:13

2k elements is a lot, but it might happen if someone spouses the idea of namespaced maps i.p.v. nested maps. work from @U066U8JQJ on his graph stuff could lead to such cases. At Synple we had pretty big nested maps, if we were to use ns instead of nesting we would have probably reached a few hundred entries (not 2k tho)

Mno11:02:04

Yo! took me a while to join.

thomas14:02:45

but now you are here and that is what counts!!!

thomas14:02:09

(= Mno Mno) simples