Re: character-by-character in a huge file

[hardburn]

Try writing a quick program that reads the file in whatever way your main program does. Just read it in and nothing else. Keep track of the run time. Then compare it to the runtime of the real program. I bet there isn't a great deal of difference. The reason is that simply reading 3GB of data using any method is going to be quite slow.

[Anomynous Monk]

I have to disagree; that code will execute about 14 ops per character (or about 42000000000 ops total); even the overhead of perl's op dispatch loop will be huge, much less the actual code of the ops themselves. Simulating just the basic overhead with perl -we'1 for 1..14000000000' (3 ops per loop iteration) runs about 36 minutes on my system; this is substantially longer than it would take to read through 3Gb.