GCC 13 vs. Clang 17 Compiler Benchmarks, Early Clang 18 & GCC 14 Development Benchmarks

[sabriah]

The publication date is wrong, as it says:

"GCC 13 vs. Clang 17 Compiler Benchmarks, Early Clang 18 & GCC 14 Development Benchmarks
Written by Michael Larabel in Software on 28 December 2022".

Was it added manually?

[david-nk]

Is the "geometric mean" still mixing in the "lower is better" metrics without inversion? If so, it is worse than useless, as it has always been.

[Michael]

Is the "geometric mean" still mixing in the "lower is better" metrics without inversion? If so, it is worse than useless, as it has always been.

That's never been the case, it's always been inverted for HIB before calculating geo mean.

[RJVB]

I've looked at a couple of of these reports thinking (hoping in vain...) that compiler benchmarks and compiler performance results would tell me something about how the compilers perform instead of how the code they generate performs.

When you're doing lots of compilations like this it has to be possible to standardise the builds as much as possible and to add at least some rough observations about much time and resources each contestant needed to build the individual benchmarks. That's pretty relevant to the question "what compiler should I use for developing X" and shouldn't miss from this kind of article if it is supposed to be of help with such questions!