Do you optimize your code based on L2 cache size?

I came across this video a few weeks back CppCon 2014: Mike Acton "Data-Oriented Design and C++" - YouTube

Basically its on optimization of code based on l2 cache size. As a relatively new programmer, I didn’t knew about these things before(how to caches work etc). So, I wanted to ask if anyone else optimizes their code based on their targeted platform cache size? And how much effort does it generally take considering we don’t write from scratch and use Engine side code as a base?

It’s a really low-level stuff, the engine team is probably the one which worries about it the most. As for the gameplay programming and other higher level areas it’s not very relevant really. So in short, I would say not to worry.