Code Translation without exceeding token limits

I have a pretty old program written in PL/1. I want to translate it to a new programming language (e.g., C#, Java, etc.). The source code has several hundreds lines of code.

How do I split the source code so I do not hit ChatGPT’s token limits?