hardburn,
Just a minor clarifcation as your post is correct, but potentially misleading. You should check out Chromatic's reply to my post saying basically the same thing as you.

Perl does not stop reading a file at a newline and have to go back out each time it encounters a newline. It reads to a buffer and each readline is taken from that buffer. The disk is not read again until the buffer is exhausted. You are showing one way of changing the buffer size. I used a different method $/ = \65536; Depending on what data munging is going on inbetween each readline and how big the file is can determine if you should use sysreads with a specified buffer as you have shown, or change the input record seperator $/, or use combination of both.

The only reason I clarify here is because I can see some inexperienced monks think that while <FILE> {}; as being inefficient, prematurely optimize, and end up with a headache back here asking for advice.

Cheers - L~R