in reply to Minimum Hash keys allocated?

You need a huge number of keys to make pre-allocation worth it. Running this several times:
use warnings;
use strict;

use Benchmark;
my $start = new Benchmark;

my $size = 10000000;
my %hash;
keys(%hash) = $size;

for my $i (1..$size) {
    my $key = "key$i";
    $hash{$key} = $key;
}

my $end = new Benchmark;
my $diff = timediff ($end, $start);
print "The code ran for \n",timestr($diff),"\n";

[download]
With 10,000,000 keys, typical numbers for pre-allocated hash keys was:
19 wallclock secs (17.36 usr +  0.94 sys = 18.30 CPU)

[download]
Typical numbers with the pre-allocation commented out:
21 wallclock secs (21.23 usr +  0.63 sys = 21.86 CPU)

[download]

I had to use 100,000 keys to even measure a difference, which was typically 0.03 seconds.

It ain't worth it.

Replies are listed 'Best First'.
Re^2: Minimum Hash keys allocated?
by flexvault (Monsignor) on Sep 27, 2010 at 11:01 UTC

    I can't show a simple script to defend my earlier claim, since the times were taken using Time::HiRes qw(gettimeofday).

    The sub-system used is 20,000+ lines of code running persistently with pre-forked servers that dynamically expand/contract on activity. The servers respond to tcp calls from a web server with a very small mod-perl cgi script preloaded.

    The design goal was to support 3 transactions per second per core with enough memory to prevent paging.

    One bottleneck was to reload all hashes required for each transaction. Testing found that pre-allocting the hash key number had a significant improvement in production.

    The last stress test, we exceeded 3.52 tranactions per second per core. However, the difference between hardware, operating systems, web-servers, mod-perl, and versions of perl is significant, so your mileage may vary!

    Thank you

[reply]

In Section Seekers of Perl Wisdom