Is there overhead by file_get_contents with offset and length vs splitting the original file and reading those single files when needed?

Question

I have a very large file (50 GByte), that I could either split in many single 2 MByte chunk files, or that I could access with file_get_contents using offsets and length of 2 Mbyte, where the used offsets are not neccessarily continuous.

So I wonder how much overhead file_get_contents does have here, or if it is even faster because the file handle is already open after the first access?

I'm using php 7.3.8 on Win10.