me, the concept of branch prediction states that the cpu predicts branches or groups of instructions likely to be executed and checks if it right, and if it's right, it buffers the instructions. How is that efficient enough, if it has to check if it's right every time, wont that be the same without branch prediction or even faster without? if i'm wrong someone help me understand
(in reply to) Way too tired to answer, I've had exactly 1.5h of sleep tonight
You'll only get a delay if the prediction is wrong.
https://www.raspberrypi.org/blog/why-raspberry-pi-isnt-vulnerable-to-spectre-or-meltdown/ Don't be fooled by the title, very comprehensive guide imho
Обсуждают сегодня