<$filehandle> also works. But in binary mode, reading large chunks of data may be more efficient.
You are not interested in parsing the file line by line, and newline chars may be between very small groups of chars, so that would force you to do more disk accesses.