Re: [ClojureScript] Re: Clojurescript string equivalence in Chrome
Issue added to v8: https://bugs.chromium.org/p/v8/issues/detail?id=4773 On Sunday, February 21, 2016 at 2:21:12 PM UTC-6, Stephen Nelson wrote: > Great job finding and analysing that. I think the appropriate place to report > upstream is https://bugs.chromium.org/p/v8/issues/list. Seeing as you did the > hard work of finding a javascript reproduction you should submit it and get > the credit, but I can if you don't want to follow up on it. > > > On Fri, Feb 19, 2016 at 4:47 AM, Francis Avila wrote: > At http://dev.clojure.org/jira/browse/CLJS-1574 I added a smaller > reproducible case that is pure JS. Basically all you need to do is use > triple-equal with mixed types. The first non-string compare will cause > subsequent string compares to be slow. If you use == instead there is no > slowdown. Clearly a chrome/v8 issue. > > > > This code isn't hot, so it's not related to the higher-tier optimizing JITs. > > > > > > On Tuesday, February 16, 2016 at 7:58:51 PM UTC-6, Stephen Nelson wrote: > > > I've had an interesting day debugging a very strange performance problem in > > Google Chrome. > > > > > > http://dev.clojure.org/jira/browse/CLJS-1574 > > > > > > `` > > > > (test) > > > cljs equiv: 0.005 seconds > > > > (= :added :ns) > > > false > > > > (test) > > > cljs equiv: 1.517 seconds > > > ``` > > > > > > I've never heard of equality causing side-effects before. My suspicion is a > > Chrome JIT bug. Has anyone seen anything like this? > > > > -- > > 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 clojurescrip...@googlegroups.com. > > To post to this group, send email to clojur...@googlegroups.com. > > Visit this group at https://groups.google.com/group/clojurescript. -- 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.
Re: [ClojureScript] Re: Clojurescript string equivalence in Chrome
Great job finding and analysing that. I think the appropriate place to report upstream is https://bugs.chromium.org/p/v8/issues/list. Seeing as you did the hard work of finding a javascript reproduction you should submit it and get the credit, but I can if you don't want to follow up on it. On Fri, Feb 19, 2016 at 4:47 AM, Francis Avila wrote: > At http://dev.clojure.org/jira/browse/CLJS-1574 I added a smaller > reproducible case that is pure JS. Basically all you need to do is use > triple-equal with mixed types. The first non-string compare will cause > subsequent string compares to be slow. If you use == instead there is no > slowdown. Clearly a chrome/v8 issue. > > This code isn't hot, so it's not related to the higher-tier optimizing > JITs. > > On Tuesday, February 16, 2016 at 7:58:51 PM UTC-6, Stephen Nelson wrote: > > I've had an interesting day debugging a very strange performance problem > in Google Chrome. > > > > http://dev.clojure.org/jira/browse/CLJS-1574 > > > > `` > > > (test) > > cljs equiv: 0.005 seconds > > > (= :added :ns) > > false > > > (test) > > cljs equiv: 1.517 seconds > > ``` > > > > I've never heard of equality causing side-effects before. My suspicion > is a Chrome JIT bug. Has anyone seen anything like this? > > -- > 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. > -- 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.
[ClojureScript] Re: Clojurescript string equivalence in Chrome
At http://dev.clojure.org/jira/browse/CLJS-1574 I added a smaller reproducible case that is pure JS. Basically all you need to do is use triple-equal with mixed types. The first non-string compare will cause subsequent string compares to be slow. If you use == instead there is no slowdown. Clearly a chrome/v8 issue. This code isn't hot, so it's not related to the higher-tier optimizing JITs. On Tuesday, February 16, 2016 at 7:58:51 PM UTC-6, Stephen Nelson wrote: > I've had an interesting day debugging a very strange performance problem in > Google Chrome. > > http://dev.clojure.org/jira/browse/CLJS-1574 > > `` > > (test) > cljs equiv: 0.005 seconds > > (= :added :ns) > false > > (test) > cljs equiv: 1.517 seconds > ``` > > I've never heard of equality causing side-effects before. My suspicion is a > Chrome JIT bug. Has anyone seen anything like this? -- 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.
