Hey, the .cirru code looks quite foreign but I guess I can kinda see what is going on. cljs$core$next is probably not what is slow here, next when working on a lazy seq will realize the next entry. So while next is fast the actual work that happens is "slow". Flame Graphs are generally better for understanding what is going on.
Anyways, some general recommendations: - do not profile in dev mode. compile with :advanced mode. use :pseudo-names true so you can still kinda see what code is running. At the very least compile with :static-fns true. - do not use partial. quamolit.alias has a bunch which I assume are called frequently. - do not use concat. Too lazy, too many allocations. - try transducers as well, they allocate less. HTH, /thomas On Thursday, April 21, 2016 at 5:02:28 AM UTC+2, Jiyin Yiyong wrote: > I got an app with ClojureScript and it's quite slow > http://repo.tiye.me/Quamolit/quamolit/ > > My code is doing a lot of rendering work, I haven't yet finished designing > the caching solution, but the CPU profile showed there's performance issue in > `cljs$core$next`. Why is that slow, did I use it wrong? > > My code is at http://github.com/Quamolit/quamolit if you are interested run > `boot compile-cirru` to get ClojureScript code, just another syntax... by > running `boot dev` to start development environment. Since I used Hit Regions > APIs, you need to turn on "Experimental Canvas features" to run the app, > check here > https://developer.mozilla.org/en-US/docs/Web/API/CanvasRenderingContext2D/addHitRegion -- Note that posts from new members are moderated - please be patient with your first post. --- You received this message because you are subscribed to the Google Groups "ClojureScript" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojurescript+unsubscr...@googlegroups.com. To post to this group, send email to clojurescript@googlegroups.com. Visit this group at https://groups.google.com/group/clojurescript.
