at first it will be a bit faster if you dont use an alias for $_. perhaps try map to increase the speed.. i´m intrested too in this case.. probably there is a solution via hashing.. u need exactly match||no match prints? or you "only" need the matched lines? <= if so it can be done faster 4 sure..

for(@test_array){print "MATCHED\n" if $_ =~ $test3 && next;
print "NO MATCH\n";} #untested
[download]

futhermore check:
How do I optimize a regular expression? and
Perl regexp matching is slow??