Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I still remember benchmarking the various optimization options in GCC and the only one that consistently and significantly improved performance on real code was -fomit-frame-pointer.


It saves one push per function call, which probably helps more than freeing up the register.


The improvement was dramatic. IIRC 20-25% faster code pretty much every time. Nothing else in the optimization set even came close.


Back in the days before on-chip caches register memory was vastly faster than main memory.


My custom builds of MySQL and PHP back in the day did exactly that.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: