I would agree with your father. "if a coder does not consider performance in their code even at the micro level, they are not good programmers." The key is to "consider performance". That is not equivalent to "do micro-optimizations at every step". I agree with most of the other comments - what used to make C programs faster may not do so today - but there are other serious things to consider: Should I use C or C++? Classes with simple overloaded operators can kill performance if you use them a lot. Does that matter? Depends on your application, but if you don't even CONSIDER it, you're not a very good programmer. Some folks might consider it for about 2 milliseconds and dismiss it, but I think far too many literally don't even consider it.

The reality is that with optimization, code can get harder to read. Code will take longer to write. Code will be somewhere between a little faster and an order of magnitude faster (that's rare). Your customers probably won't know the difference.

Another reality is that people like to reuse code, and where it ends up may be quite different than when you wrote it. Suddenly people may care.

As an example, say you wrote something like Angry Birds on a PC, or for a game console. You disregarded optimization in your physics system and your graphics system - because 2.5GHz dual cores are basically the minimum these days, and your game works good enough. Then smart phones come along and you want to port it. Oh darn, a 600MHz ARM core?

Think of a web site - something thrown together in a weekend like hotornot. You use whatever database is handy, write everything with rapid development in mind, launch by emailing your friends a URL. 3 months later your server is dying from overload and there's nothing you can do to scale up. Happens all the time. The largest web sites have power bills measure in hundreds of kilowatts or even megawatts. Think about that, there are megawatts to be saved through code optimization.

Like your father said, you must at least consider it. Most folks don't even know how much performance is on the table and gloss over it with a quick quip about "premature optimization" and "don't do it". These are not good programmers.

This is not a popular opinion on these sites. Bye bye rep....

I would agree with your father: "If a coder does not consider performance in their code even at the micro level, they are not good programmers." The key is to "consider performance". That is not equivalent to "do micro-optimizations at every step". 

I agree with most of the other comments - what used to make C programs faster may not do so today - but there are other serious things to consider: Should I use C or C++? Classes with simple overloaded operators can kill performance if you use them a lot. Does that matter? It depends on your application, but if you don't even CONSIDER it, you're not a very good programmer. Some folks might consider it for about 2 milliseconds and dismiss it, but I think far too many literally don't even consider it.

The reality is that with optimization, code can get harder to read. Code will take longer to write. Code will be somewhere between a little faster and an order of magnitude faster (that's rare). Your customers probably won't know the difference.

Another reality is that people like to reuse code, and where it ends up may be quite different than when you wrote it. Suddenly people may care.

As an example, say you wrote something like Angry Birds on a PC, or for a game console. You disregarded optimization in your physics system and your graphics system - because 2.5 GHz dual cores are basically the minimum these days, and your game works good enough. Then smart phones come along and you want to port it. Oh darn, a 600 MHz ARM core?

Think of a web site - something thrown together in a weekend like Hot or Not. You use whatever database is handy, write everything with rapid development in mind, launch by emailing your friends a URL. Three months later your server is dying from overload and there's nothing you can do to scale up. It happens all the time. The largest web sites have power bills measured in hundreds of kilowatts or even megawatts. Think about that, there are megawatts to be saved through code optimization.

Like your father said, you must at least consider it. Most folks don't even know how much performance is on the table and gloss over it with a quick quip about "premature optimization" and "don't do it". These are not good programmers.

This is not a popular opinion on these sites. Bye bye reputation points....