I'm not suggesting that any of this applies to the O/P, in this case.

You can't blame the folks who use 'C' as a means for avoiding a good read of the datasheet for using the optimization tools as they are without study. Likewise, the common assumption seems to be that using HLL is somehow "better" than using ASM. I don't believe that, but I do believe that it's cheaper for the sponsor, and it leads to more "maintainable" code if properly documented.

Too few compiler writeups explain that optimization can, in fact, be improved upon, sometimes vastly, by manual effort.

I use 'C' as an example. This is not an indictment. Since we have tools, e.g. compilers, simulators, macro-assemblers, linkers, we should use them. However, it's not a bad idea to know their weaknesses, too.

Tools are supposed to make the work easier for us, so we can focus on making it better. Isn't that so?

RE