4 ms·

LL/SC is performant, it just doesn't scale to high core counts. The VEX encoding is actually only rarely longer than the legacy one, and frequently it is short

LL/SC is performant, it just doesn't scale to high core counts. The VEX encoding is actually only rarely longer than the legacy one, and frequently it is shorter.