This is nonsense. If s is fixed-size at compile-time, you would never use apply to begin with. Why bother with (applyn 10 + [1 2 3 4 5 6 7 8 9 10]) when you could just write (+ 1 2 3 4 5 6 7 8 9 10)?
On Sunday, October 7, 2012 2:15:28 PM UTC-7, Marc Dzaebel wrote: > > *apply *is slow. However you can increase performance by 60% with the > following macro, if you have a fixed length in S. > > (defmacro *applyn *[n f & s] > (loop [curr `(list* ~@s), n n, vars[] vals[]] > (if(pos? n) > (let[v(gensym)] > (recur v (dec n) (conj(conj vars v) (if (seq vars) (list > 'next curr) curr)) > (conj vals(list 'first v)))) > `(let[~@vars] ~(cons f (seq vals)))))) > > (let[t(fn[](*apply *+ '(1 2 3 4 5 6 7 8 9 10)))] (time(dotimes [_ > 1000000] (t)))) ; ~680 msec > (let[t(fn[](*applyn *10 + '(1 2 3 4 5 6 7 8 9 10)))] (time(dotimes [_ > 1000000] (t)))) ; ~220 msec > > So, if you have inner loops, that must be optimized for performance, you > might remember this possibility. Even other functions could be optimized > this way. However, "*premature optimization* is the root of all evil". > Beside, the generated code is more space consuming. > > *Marc* > -- 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
