alt Hacker NewsI’ve spent many years working on hpc problems, and most people simply don’t have , or don’t care about, or don’t know what to do with performance problems.
Don’t have to because they usually have a full machine, and unless they saturate it, the problem doesn’t exist.
Don’t care, because many people think it’s alright if things are slow, or inefficient ( the user can wait, or will live with 100ms instead of 1ms).
Don’t know what to do - this one is interesting : many people don’t realize it’s actually possible to get the same result , but faster.
> Don’t care, because many people think it’s alright if things are slow, or inefficient ( the user can wait, or will live with 100ms instead of 1ms).
It's not that.
I've encountered code where someone wrote complicated "scale out" code instead of learning how to use a database and fix the n+1 problem; or another case where someone wrote complicated caching code instead of learning how to avoid cartesian explosion.