Mike Acton is (if memory serves) tech lead at Insomniac Games--folks behind Ratchet & Clank, the Resistance series, and others.
This is a presentation which is mean in spirit but excellent in content; just note that a lot of this is caused by their targeting of the Cell processor at the time.
Overall, there are a lot of great refactoring questions he raises, and really it's a good read if you'd like to learn what sort of mindset you'd use when developing race-horse performance-sensitive code. In general, I don't agree with some of the advice, but in their problem domain (real-time interactive graphics a.k.a. games) it makes complete sense.
This is a presentation which is mean in spirit but excellent in content; just note that a lot of this is caused by their targeting of the Cell processor at the time.
Overall, there are a lot of great refactoring questions he raises, and really it's a good read if you'd like to learn what sort of mindset you'd use when developing race-horse performance-sensitive code. In general, I don't agree with some of the advice, but in their problem domain (real-time interactive graphics a.k.a. games) it makes complete sense.
EDIT:
This slide (http://macton.smugmug.com/gallery/8936708_T6zQX#!i=593430457...) made me chuckle.