Number of CPUs + 2 is what pmap uses, and I assumed the idea was to keep all the CPUs busy in the event that one finishes before the others. I wrote it before I did testing with npmap. Since reading your last post, I did a bit of testing with modified versions of zpmap and found that it isn't making much difference on a quad-core machine. It might be a few ms faster on a 3.5s call, but it's so close that it's probably within the margin of error.
It may turn out to be a better approach to rewrite pmap instead of wrapping it. I'm wondering if the attempt to make it both semi-lazy and parallel isn't optimal for speed. -- 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
