update:
I seem to be getting different times on each run (and different winners too), but over the whole split does appear to be the winner. I've changed the regex to a more condensed version.

---

#!/usr/bin/perl -w

use Benchmark 'timethese';
use strict;

print "split: ", use_split(), "\n";
print "regex: ", use_regex(), "\n";

sub use_split {
    $_ = "  one two three    four five six";
    split;        # See Jouke's post (this may be faster). This is dep
+recated though.
    return $_[3];
}

sub use_regex {
    $_ = "  one two three    four five six";
    s/^\s+//;
    m/^(?:[^\s]+\s+){3}([^\s]+)/;
    return $1;
}

timethese(-10, {        regex => \&use_regex,
                        split => \&use_split });
timethese(-50, {        regex => \&use_regex,
                        split => \&use_split });
timethese(-100, {       regex => \&use_regex,
                        split => \&use_split });
[download]

The results (updated to 10, 50 and 100) (...drumroll please...):

Benchmark: running regex, split, each for at least 10 CPU seconds...
     regex: 12 wallclock secs (10.64 usr +  0.00 sys = 10.64 CPU) @ 40
+522.18/s (n=431156)
     split: 10 wallclock secs (10.02 usr +  0.00 sys = 10.02 CPU) @ 37
+290.82/s (n=373654)
[download]

and:

Benchmark: running regex, split, each for at least 50 CPU seconds...
     regex: 68 wallclock secs (54.11 usr +  0.13 sys = 54.24 CPU) @ 40
+957.93/s (n=2221558)
     split: 56 wallclock secs (49.95 usr +  0.05 sys = 50.00 CPU) @ 38
+128.48/s (n=1906424)
[download]

and

Benchmark: running regex, split, each for at least 100 CPU seconds...
     regex: 136 wallclock secs (110.38 usr +  0.32 sys = 110.70 CPU) @
+ 40548.93/s (n=4488766)
     split: 112 wallclock secs (100.00 usr +  0.02 sys = 100.02 CPU) @
+ 38995.86/s (n=3900366)
[download]

[~]

---

---

• Are you posting in the right place? Check out Where do I post X? to know for sure.
• Posts may use any of the Perl Monks Approved HTML tags. Currently these include the following:

  <code> <a> <b> <big> <blockquote> <br /> <dd> <dl> <dt> <em> <font> <h1> <h2> <h3> <h4> <h5> <h6> <hr /> <i> <li> <nbsp> <ol> <p> <small> <strike> <strong> <sub> <sup> <table> <td> <th> <tr> <tt> <u> <ul>

• Snippets of code should be wrapped in <code> tags not <pre> tags. In fact, <pre> tags should generally be avoided. If they must be used, extreme care should be taken to ensure that their contents do not have long lines (<70 chars), in order to prevent horizontal scrolling (and possible janitor intervention).
• Want more info? How to link or How to display code and escape characters are good places to start.