Date: Thu, 13 Sep 2001 14:09:51 GMT
From: tadmc@augustmail.com (Tad McClellan)
Subject: Re: How to preallocate memory for my perl script (UNIX)
Message-Id: <slrn9q1ckj.eec.tadmc@tadmc26.august.net>

Stan Brown <stanb@panix.com> wrote:
>I have a perl script that retrieves a fairly large amount of data from
>Oracle. I'm trying to speed it up.
>
>I have obsered that the first time I do a lareg fetch, and store it in a
>perl variable, 


What kind of variable? Array? Hash?


>the process size of the script slowy increases. Now ehen I
>do a second query, it runs _much_ faster.
>
>I beleive that a part of this is because perl has to make many small
>requests to the OS for emeory, the first time. So I would like to do
>something, before the first fetch to get perl to request a large chunk of
>memory from the OS, the undefine whatever is using this memeory, thus
>making that memory available to perl for reuse.
>
>Sugestions?


You can preallocate memory:

   my @array;
   $#array = 1_000_000;  # a million elements
or
   my %hash;
   keys %hash = 1_000_000;  # a million buckets (perldata.pod)


And empty it out when you are done:

   @array=();  # keep the large memory associated with @array 'cause
               # you're going to use it again

   undef @array;   # return memory to perl process, no longer associated
                   # with the array, will need to realloc when you
                   # reuse the same array (so don't do this)

-- 
    Tad McClellan                          SGML consulting
    tadmc@augustmail.com                   Perl programming
    Fort Worth, Texas