You could try it with word boundaries:

my @stuff = grep /\b run \b/x, @other_stuff;
[download]

grep /\brun\b/,@array;
[download]

print grep { m{\brun\b} } qw/run runnable running run awordendingwithr
+un/;
[download]

Output:

runrun
[download]

See perlre.

Note that the answers above rely on the array being comprised of single words (which may well be the case), However, if the elements being searched contain multi-word phrases or (for example) sentences (which is not specified in the OP), the results may not be as illustrated.

#! /usr/bin/perl
use strict;
use warnings;

my @other_stuff = ("run", "running", "runs", "runt", "I run a mile per
+ day.", "Run, rabbit, run!", "One mile per day is easily runnable");

for my $element(@other_stuff) { print $element ."\n"; }

print "-" x20 . "\n";

my @stuff = grep /\brun\b/x, @other_stuff;

for (@stuff) { print $_ . "\n"; }
[download]

produces...

ww@GIG:~/pl_test$ perl greprun.pl
run
running
runs
runt
I run a mile per day.
Run, rabbit, run!
One mile per day is easily runnable
--------------------
run
I run a mile per day.
Run, rabbit, run!
ww@GIG:~/pl_test$
[download]

Precise specification of a problem-area makes it easier for Perl Monks to provide proper answers (and standard spelling is a "good thing," too).

is your friend.

Thank you for all of you who reply this question. :)

One more question on the word boundary: I read the documentation and the document said,

"is a spot between two characters that has a \w on one side of it and a \W on the other side of it (in either order), counting the imaginary characters off the beginning and end of the string as matching a \W "

Can I said the word boundary is any character in class \w?(e,g, space, tab, newline?)

Can I said the word boundary is any character in class \w?(e,g, space, tab, newline?)

Not really, no. To illustrate the difference:

my $string = 'run rabbit run';
print "Matched with word boundary\n" if $string =~ /\brun\b/;
print "Matched with non-word char\n" if $string =~ /\Wrun\W/;
[download]

Update: I (and the above code) assume you meant 'class \W' rather than 'class \w.' If you meant 'class \s', the answer is the same:

print "Matched with whitespace\n" if $string =~ /\srun\s/;

The \w character class does not include space, tab or newline.   The \w character class is comprised of letters, numbers and '_' the underscore character.

Also, the \b word boundary does not match a character, it is a zero-width assertion, it matches between characters.