Stolen Adapted from from Unix, it's been around since the early days of DOS. It allows you to compare two files, to see if they're identical, and if not, show the differences.
In its early days, it was ideal to see if that copy of a file onto a floppy was accurate. Now, I sometimes use it to see if the copy onto a CD-R is ok.
How do you think it – the version with Windows 2000 anyway – works? You've got two 700M byte files. How would you compare them?
Read them in small chunks, and compare those?
Or read one entire file into memory, all 700M of it… Even with over a gigabyte of RAM, that results in more hard disk activity as the memory is 'paged' to and from disk.
Yep, it's the latter. At least that's the only reason I can think of for its memory requirements slowly climbing to just over the size of the file.