On Sat, Jan 3, 2009 at 5:50 AM, Jason <jawo...@berkeley.edu> wrote:
>
> As you can see here, accessing elements of nested subvectors is slow.
>
> user> (doseq [i (map #(Math/pow 2 %) (range 0 16))]
> (print i " ")
> (let [sub (nth (iterate #(subvec % 0 1) [1]) i)]
> (time (dotimes [_ 10000] (sub 0)))))
> 1.0 "Elapsed time: 1.929 msecs"
> ...
> 128.0 "Elapsed time: 7.477 msecs"
> ...
> 32768.0 "Elapsed time: 10342.465 msecs"
>
> It seems the speed hit can be avoided by adding a special case to the
> SubVector constructor in APersistentVector.java:
>
> public SubVector(IPersistentMap meta, IPersistentVector v, int start,
> int end){
> super(meta);
> if (v instanceof SubVector) {
> SubVector s = (SubVector)v
> this.v = s.v;
> this.start = start + s.start;
> this.end = end + s.start;
> } else {
> this.v = v;
> this.start = start;
> this.end = end;
> }
> }
>
> I've tested this a bit and it appears to work properly. Does this
> seem like a good idea?
It seems like a good idea to me.
Have you sent in your CA? See: http://clojure.org/contributing
--Chouser
--~--~---------~--~----~------------~-------~--~----~
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
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
-~----------~----~----~----~------~----~------~--~---
