I just explained to a coworker why uint16 vector constants >=0x8000 in our code are written “0xbaad – 0x10000” (because the official Intel intrinsics have initializer funcs for int16 but not uint16 and passing uint16s gives warnings) and I remembered again how much this sucks
— Fabian Giesen (@rygorous)
Aug 19, 2021
from http://twitter.com/rygorous
via IFTTT