On Tue, 2025-05-27 at 14:56 +0200, Christer Solskogen via Gcc-help wrote: > Hi! > > https://gcc.gnu.org/install/configure.html claims that > --enable-host-shared yields a slightly slower compiler, but has there > been any benchmark to backup that claim? Or how big of an impact it > has? Ten years ago it was often a few percents on x86 (32 bits) or x86-64. (of CPU time, in seconds) If I remember correctly in practice it relates to the fact that on some processors you need (at ABI level) to use one more processor register in some (not all) parts of the generated code. And current processors all have more registers and more cache, so the effect could be less. Also GCC improved a lot since then. But you need to measure of course, if you want precise measures Of course it depends upon the compiled code. You will find exceptions where that matter more. Regards. -- Basile STARYNKEVITCH <basile@xxxxxxxxxxxxxxxxx> 8 rue de la Faïencerie http://starynkevitch.net/Basile/ 92340 Bourg-la-Reine https://github.com/bstarynk France https://github.com/RefPerSys/RefPerSys
