Benchmarks

[FireBurn]

trafficstars

I'm not sure if you noticed the phoronix article: https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.19-O3-March-Native

Which showed -march=native being much slower than just using -O2 or -O3

Has anyone else managed to verify this? Was it a fluke on that particular processor or arch? Was it a GCC 11 issue? etc etc

[aliceinwire]

oh this looks interesting. looks like is thanks to this patch series: https://lore.kernel.org/lkml/[email protected]/