He's very data and performance oriented. He gives advice on how to write games, high-performance systems. His advice is solid and great even, but not when your source of latency is a cloud DB.
I beg to differ even on that. When your source of latency is a cloud DB, learning about CPU caches won't help you, but learning to think about end-to-end performance absolutely will. "What sequence of queries achieves my goal with the shortest critical path?" is very similar to "What sequence of vector instructions achieves my goal with the shortest loop-carried dependency?"
> but learning to think about end-to-end performance absolutely will.
In a cloud DB webservice what will save you is having a good tracing/debugging story. From it, it will be easy to see E2E performance. Sure, some things will translate, but going SIMD optimized layout on a class, that will have to talk to DB across the ocean is a fool's errands.
To optimize you must have a good target and make sure you are measuring correctly.
Why do you think that making a minimum number of queries to a cloud database is anything other than conceptually similar to making a SIMD optimized layout on a class?
That's not what I was talking about. Optimizing queries to DBs is nothing like optimizing SIMD layout.