> More findings: The reason that the Clojure's original sort is 8 times slower
I don’t see that on my machine. I’m running 1.1.0-master-SNAPSHOT with
Apple’s Java 6 VM in case that has anything to do with it, but here's
what I get (after running the tests several times to warm up hotspot):
user=> (def v (vec (take 10000 (repeatedly #(rand-int 100000)))))
user=> (time (dotimes [_ 1000] (sort v)))
"Elapsed time: 4376.471 msecs"
user=> (defn sorted-vec [coll]
(let [a (into-array coll)]
(java.util.Arrays/sort a)
(vec a)))
user=> (time (dotimes [_ 1000] (sorted-vec v)))
"Elapsed time: 3254.371 msecs"
user=> (defn sorted-vec-2 [coll]
(let [a (to-array coll)]
(java.util.Arrays/sort a)
(vec a)))
user=> (time (dotimes [_ 1000] (sorted-vec-2 v)))
"Elapsed time: 2599.63 msecs"
So sorted-vec is faster, but not an order of magnitude, and sorted-
vec-2 is faster again.
Another alternative that may be worth considering is leaving the data
in the array and using aget to access elements (this should give you O
(1) access times vs. O(log32N) AFAIK). This may be a solution if
you're not mutating the data in the array, but I'd be careful about
this optimisation unless it really gets a large speed boost for your
code.
--
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
