On Sat, Apr 27, 2013 at 11:36 PM, Lee Spector <lspec...@hampshire.edu>wrote:
> > I don't want to be too much of a downer, but our experience running > Clojure on 48-core machines was pretty depressing, with our speedups often > being near zero or even negative (that is, running on 48 cores can be > slower than running on 1, for a task that would seem to be a good candidate > for parallelization, and with what would seem to be good techniques for > doing the parallelization). Several people tried to help us diagnose/fix > the problems (see this thread: > https://groups.google.com/forum/?fromgroups=#!topic/clojure/48W2eff3caU), > but I don't think that there has been a definitive resolution. > The resolution is obvious, isn't it? The JVM needs to evolve to have thread-local polymorphic-call-site type profiles, analogous to the already-existing thread-local allocation buffers used to avoid allocation contention across threads, and thread-local copies of native methods. -- -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
