Comment by jmward01

I have almost always found that simple code runs faster than complex code. I think this is because optimization is likely an NP problem and like all NP problems, the best algorithm we have for solving it is divide and conquer. The core thing about D&C is that you divide until you reach a level that you can actually find the optimum answer within the resources given but accept that by dividing the problem you will likely have some high level inefficiencies. This means simple code, code you understand all pieces of, can actually be locally optimum. When I see people try to optimize code I often see them fall into that trap of optimizing too large/complex a problem which leads to them not actually being able to find a true local optimum and, often, making far slower code than had they just tried for much simpler code. This likely NP behavior runs rampant in software development where we often think we can design some elaborate process to design things and when things fail it was because people failed to follow the process and not because the problem was NP. We all love building machines which is why we likely do this, but unless you know something the rest of the world doesn't then D&C, and admitting you are only looking for a local optimum, is the only algorithm we have for attacking NP problems. (Shameless plug here for smaller, far more autonomous teams instead of monolithic dev shops with one big shared feature list)