For those interested, I managed to improve the performance of my
original program from 2 minutes 40 seconds to decode 1000+ files down
to 2 minutes.  I'm still far from my goal, but it's an improvement,
especially since the code is shorter and (IMO) cleaner.  You can see
it here: 
http://bitbucket.org/gnuvince/clj-starcraft/src/tip/src/starcraft/replay/parse.clj

And here's another question, running the program with -Xprof shows
that nearly 20% of my execution time is spent calling
java.lang.reflect.Array.set.  Is there something wrong with the way I
type hint my array in make-reader?

Thanks,

Vincent.

On Mar 19, 8:12 pm, Vincent Foley <vfo...@gmail.com> wrote:
> Hello,
>
> For the past few days, I've been trying, unsuccessfully, to make an
> application I wrote faster.  A Java program that performs, more or
> less, the same task takes 12 seconds (on my machine) to parse 1000
> files; my Clojure program takes nearly 3 minutes.  This more than an
> order of magnitude slower!  Using the profiling tools available with
> the JVM, I quickly determined which function was the costliest.  I
> copied it into a simple script file to profile it in isolation.  I
> have made the script and the profile results (long!) available at this
> URL:http://gist.github.com/82136
>
> I'm finding the results puzzling: is dereferencing a var *that*
> expensive?  Can anyone tell me if they see something fundamentally
> wrong with my approach that would explain this abysmal performance?
>
> Thank you,
>
> Vincent.
>
> P.S.: I am using Sun's JVM 1.6.0_10 as shipped in Ubuntu Ibex.  My
> machine is an Athlon 64 X2 4200+ with 3 GB of RAM.
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---

Reply via email to