Attached: pepedevil_by_rpglord-dbebq3y.png (400x400, 9K)
>compile linux & drivers >with -Ofast -march=native
Owen Miller
Other urls found in this thread:
lemire.me
twitter.com
Austin James
why is that /devillish/ ? I'm pretty sure that's the right way to do it if you don't wish to wait ages
Isaiah Brown
lmao retard
vectorization cannot be used in the kernel
William Perez
Is this what you guys mean by devs?
Cameron Martin
>he doesn't know it's already -may
Oliver Wilson
>php artisan config:cache
William James
Watch even senior devs not understanding why the database is STILL pointing to homestead
Luke King
>-g -O0 -fno-inline
Nolan Hernandez
have some patience instead of risking yourself for a massive fuckup
Adrian Reyes
>-ffast-math
Oliver Kelly
I always hear this but have never had bugs appear. What could possibly go wrong?
Angel Sullivan
anything. it's easily the most important thing you need to compile so you really shouldn't take any risk
Jack Diaz
>aka -Ointel
Brody Lee
gcc -march=k12 -Intoddlers_btfo -Deeznuts
Andrew Sullivan
Oh mama
Andrew Cook
wait is that true?
Just changing compile flags can fuck up the kernel?
>The absolute state of C & UB
Thomas Richardson
It can't, GCC heuristics have gotten incredible since 3.x and 4.x when everything really would go to shit
Hudson Cooper
Hey it actually works pretty well
Easton Walker
-O3 -march=native the world
Adam Reyes
Sometimes -march=native has worse options than -march=, output gcc's verbose flags and see the diff
Cooper Brown
native takes only those that are supported by the system
Ian Sullivan
Use -march=native -mtune=native then.
Nolan Powell
Why don't I feel any cconsequences?
Matthew Morales
nope, march=native implies mtune=native, you can google it ;-)
Joshua Rivera
Yes, no, and sort of.
lemire.me
Justin Nguyen
read the comments below the post, it was a bug of an old gcc af version, which got fixed
Brayden Ross
>still compiling in 2019
lol?