On 01/05/2013 07:19 PM, Charles Oliver Nutter wrote: > Hello all! > > I can accept that invalidating a SwitchPoint that's in use would cause > some perf impact. But I'm wondering why we'd see the same perf hit for > invalidating a SwitchPoint that has never been bound. Is there a > reason for this? > > I have a few places in JRuby where I eventually give up on caching > certain values if they fail too many times, but on the value updating > side I'm still invalidating SwitchPoints. If they're never actually > used other than to invalidate, shouldn't invalidating be cheap?
No. invalidating a SwitchPoint that is JITed cost your an harm, invalidating a SwitchPoint which is not JITed is just a volatile write but you still have to update the callsite to use a new SwitchPoint. > > - Charlie Rémi _______________________________________________ mlvm-dev mailing list mlvm-dev@openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
