asterite has joined #crystal-lang
psychopat has joined #crystal-lang
e_dub has joined #crystal-lang
kostya has joined #crystal-lang
kostya has joined #crystal-lang
waj has joined #crystal-lang
asterite has joined #crystal-lang
<kostya> hi
<kostya> i experiencing strange perfomance: https://gist.github.com/kostya/c1748e5dde4aefa4cb14
<kostya> asterite, any ideas?
<asterite> Let me see...
e_dub has joined #crystal-lang
<asterite> @kostya I have no idea… the generate ll code is very different if you add that third yield
<kostya> :)
waj has joined #crystal-lang
<kostya> maybe excess limit for inline
<asterite> Maybe
<asterite> Probably some llvm optimization thing
<asterite> Blocks are always inlined (by us), I don't know if llvm is smart enough to extract that to a function
<asterite> but if that's the case, maybe that's why it becomes slow
<kostya> how to generate ll code?
<asterite> You can use the -ll flag
<asterite> it generates it inside a .crystal directory
<asterite> Normally you get one .ll for each class (and one for the main program), but if you compile with —release only one is generated, main.ll
travis-ci has joined #crystal-lang
travis-ci has left #crystal-lang [#crystal-lang]
<travis-ci> [travis-ci] Build details : http://travis-ci.org/manastech/crystal/builds/20356961
<travis-ci> [travis-ci] manastech/crystal#946 (master - cad716f : Ary Borenszweig): The build passed.
waj has joined #crystal-lang
waj has joined #crystal-lang
kukushnik has joined #crystal-lang