bigbot

I don't think anyone here would tell you that a perl script could/would be faster than a "highly" optimized C program ( I could be wrong ). And there is nothing wrong with using the system command 'grep' to produce your solution.

But if you still want to improve the execution time, some things to look at:

• In your test script you used the regex match
  $matchCount++ if ($_ =~ /$string/o);

  however in this case, a simple "index" test should be faster

• If you read blocks of data ( read or sysread ) you would decrease the difference between grep and perl, but now the complexity of you script has increased.

All of these improvements are nice, but is it worth spending the time to debug the improved script to gain 'nn' seconds at execution time. That is your call -- balancing your time efficiency versus the script efficiency. Chapter 24 of The Camel book does an excellent job of explaining the trade-offs.

Good Luck