Author's only analysis is on the size of binaries. And while they are right to the extent of the point they are making, size is only one dimension. Every compi

Author's only analysis is on the size of binaries. And while they are right to the extent of the point they are making, size is only one dimension. Every compiler, but especially optimizing compilers, makes hundreds of tradeoffs as it generates object code. Some for speed of compilation, some for speed of resultant code, some for size. This is true even at various optimization levels. They all have thresholds and decision points about what makes better code, given the hundreds of dimensions they can be measured against. The only true answer to this question is to build your own code with a variety of compilers, and see which produces the results that makes the best trade offs for your own project. Fastest running, fastest build, smallest code, smallest footprint, most standards conforming, whatever else. Your own code and needs are what governs.