Just heard of a netlist extraction script that took 3 hours to complete. If you’re not familiar, we’re talking about removing 10k or so of text from a 50-100 megs of text. Should this take 3 hours?
Lets see, search all the lines for the module in question, remove lines until the end of that module.
No, that should not take 3 hours. So why does it? The answer we got is that an intern wrote the script and its performance, due to algorithm choice, degrades exponentially with file size. Huh? How did that get released to the general population? Sure it works, but 3 hours for something that should take minutes is a glaring error in algorithm choice.
This tells me that this intern’s mentor didn’t do a good job. Mistakes in algorithm choice are common with green engineers. Mentors need to look out for that and need to pay close attention to an intern’s work. Algorithm review, code review are all required to ensure a good final result.
In this case the result was not good. I blame the mentor.
hopefully I did a good job when you were my mentor :-)
ReplyDelete