Reading a 30Mb file one time instead of 500 times should be quicker. How about posting what you tried? Maybe there is something you missed in your implementation.