On Aug 9, 6:08 am, Nicolas Oury <nicolas.o...@gmail.com> wrote: > > If I do my pmaptest with a very large Integer (inc 2000000000) instead > > of (inc 0), it is as slow as the double version. My question is, > > whether Clojure may has a special handling for small integers? Like > > using primitives for small ints and doing a new Integer for larger > > ones? > > It seems a confirmation that it is a memory allocation related problem. > (I think for some reason, this is store as a long, the same size as a > double). > > This can means two things: > - as I hypothesized before, we are measuring the memory bandwidth > - we fill the Thread Local Heap so fast, that there is a lock > contention for the common heap. > > I don't know how to separate one from the another... You could write a task that doesn't allocate, but does hammer on memory bandwidth & have this run in parallel at the same time. You'd need an array that is larger than your cache, and then stride through it at cache line size steps. If the runtime doesn't decrease then you are seeing lock contention.
Cheers, Brad --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
