Try this:
(def *valid-chars* [ \a \b \c \d \e \f \g \h \i \j \k \l \m
\n \o \p \q \r \s \t \u \v \w \x \u \z
\0 \1 \2 \3 \4 \5 \6 \7 \8 \9 ] )
(defn generate-key [keylength]
(for [x (range keylength)] (nth *valid-chars* (rand-int (count *valid-
chars*)))))
(defn generate-random-keys [numkeys keylength]
(for [x (range numkeys)] (generate-key keylength)))
(time (into [] (generate-random-keys 100000 1024))) ;; 100k strings,
1024-chars each
I get "Elapsed time: 384.562 msecs". (MacPro/2.66Ghz)
I didn't check the memory usage and it may not be the most idiomatic
clojure but it's a bit easier to read and understand.
On Oct 30, 3:56 pm, Chick Corea <chick.zco...@gmail.com> wrote:
> How do I make this code faster? And leaner, i.e., use less memory ?
>
> It's a simple function to generate NUMKEYS strings of length
> KEYLENGTH,
> store them in a Java array then store them in a HashMap<String,Object>
> hash-map. [it doesn't store them in the hash-map yet; but it
> allocates it]
>
> (set! *warn-on-reflection* true)
>
> (def *valid-chars* [ \a \b \c \d \e \f \g \h \i \j \k \l \m
> \n \o \p \q \r \s \t \u \v \w \x \u \z
> \0 \1 \2 \3 \4 \5 \6 \7 \8 \9 ] )
>
> (defn generate-random-keys [ numkeys keylength ]
> (time
> (let [ rand (new java.util.Random)
> key (make-array Character/TYPE keylength) ;; :initial-
> element \a))
> keys (make-array String numkeys)
> data (new java.util.HashMap (int numkeys))
> ]
> (.println System/out (format "generating [%d] keys of length
> [%d]" numkeys keylength ))
> (dotimes [i numkeys]
> (dotimes [j keylength]
> (aset key j (nth *valid-chars* (. rand nextInt (int 36)))))
> (let [tmp (String. #^chars key) ]
> ;; (.println System/out (format "keys[%d] => [%s]" i
> tmp))
> (aset keys i tmp)))
> (.println System/out (format "inserting [%d] keys into HashMap"
> numkeys)))))
>
> (generate-random-keys 100000 1024) ;; 100k strings, 1024-chars each
>
> The corresponding Java code (w/ the hash-insert omitted in the clojure
> version)
> runs in 5.5sec and uses 290MB.
>
> This code runs (which omits the hash-insert) runs in 17.8sec and uses
> 353MB.
>
> I thought that I added all of the casts and "warn-on-reflections" that
> would help.
> Also, I thought using Java data structures directly would bypass some
> of Clojure's
> overhead. What am I missing?
>
> Granted, the initial memory allocated is
>
> * for POJ, ~12MB
> * for Clojure, ~90MB
>
> That may or may not be relevant. It's hard to know w/ auto mem-mgmt.
>
> CHICKEE
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---
