Beefy Boxes and Bandwidth Generously Provided by pair Networks
P is for Practical

Re: Fast data structure..!!!

by samtregar (Abbot)
on Apr 15, 2008 at 16:34 UTC ( #680566=note: print w/replies, xml ) Need Help??

in reply to Fast data structure..!!!

This is around 20% faster in my tests and should use less memory (I didn't check how much):

my %save; for(0 .. 2000000) { $save{$_,$_+1}=$_+2; }

The difference is that this version creates just one hash and fills it with data in compound keys (joined with $;, \034 by default). Your version created lots of sub-hashes, allocating memory for each one. Using an array is even faster though.

I once solved a problem like this by using XS to compile a shared object containing a packed C data structure with my data, along with accessor routines also written in C. Then at runtime the Perl code just loaded the shared-object to get access to the data. It was really, really fast but I doubt I'd do it again - Berkeley DB has pretty much solved this problem.


Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://680566]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others scrutinizing the Monastery: (8)
As of 2016-10-21 10:00 GMT
Find Nodes?
    Voting Booth?
    How many different varieties (color, size, etc) of socks do you have in your sock drawer?

    Results (288 votes). Check out past polls.