I never replied to this! Sorry! Threadpooled mode does improve the spinup time of threads under JRuby, but I've realized lately most of that benefit could be realized by pooling ThreadContext instances instead. The threadpool implementation for Ruby threads just seems to introduce complexity, and I've never really felt 100% confident in it.
If we removed it in 1.6, would it bother you or anyone else? On Fri, Apr 30, 2010 at 4:22 PM, Uwe Kubosch <u...@datek.no> wrote: > We fire quite a few (one or two per second) short-lived threads in our thick > clients, and we are planning to try thread-pooled mode in a week or so. The > description of thread-pooled mode promises performance improvement when using > short-lived threads. Still true? > > > Uwe > > On Apr 23, 2010, at 2:57 PM, Charles Oliver Nutter wrote: > >> Looks like the improvements I committed last night to eliminate >> RubyThread leakage have caused the threadpool mode to break: >> >> http://ci.jruby.org/job/jruby-test-all/126/console >> >> I'll look at fixing that today, but I have to ask: is anyone using or >> interested in using threadpooled mode? I added it a long time ago >> intending to make it solid for all uses, but I don't think anyone has >> started using it during those intervening years. Maybe it's more >> trouble than it's worth? >> >> - Charlie >> >> --------------------------------------------------------------------- >> To unsubscribe from this list, please visit: >> >> http://xircles.codehaus.org/manage_email >> >> > > -- > Uwe Kubosch > Systems Developer > Datek Wireless AS > u...@datek.no > http://datek.no/ > > > > > > --------------------------------------------------------------------- > To unsubscribe from this list, please visit: > > http://xircles.codehaus.org/manage_email > > > --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email
