On Mon, Dec 28, 2009 at 9:39 PM, Timothy Pratley <timothyprat...@gmail.com> wrote: > I find this quite interesting because Meikel has effectively created a > faster version of reductions: > > (defn cumulative > ([accum-fn s] > (rest (cumulative accum-fn 0 s))) > ([accum-fn accum s] > (lazy-seq > (when-let [[f & r] (seq s)] > (cons accum (cumulative accum-fn (accum-fn accum f) r)))))) > > (defn left-total > [coll] > (map list coll (cumulative + 0 coll))) > > I did some microbenchmarking and it does appear to be 2x as fast as > reductions, and processes the special case left-total just as fast as > one of the original proposals. Oddly left-tot2 performs very poorly > though it was claimed to be fast. I suppose this could be yet another > case of microbenchmarking being evil - but looking at the code the > reductions version seems to use an atom which implies some unnecessary > overhead? Or maybe I am missing something important about their > differences. > > I followed the link to the discussion about the original reductions > and it all seems to be pre-lazy-seq so maybe this is just a case of > the function should be updated? > >
This was discussed before, the new version never made it into a patch: http://groups.google.com/group/clojure/msg/43de40f078a291cc Note your use of destructuring in when-let with & makes your cumulative slightly non-lazy. Rich -- 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
