Remix.run Logo
jeffbee 14 hours ago

He isn't trying to determine how well it works. He's trying to determine how large it is.

Night_Thastus 14 hours ago | parent [-]

Their post gives the impression that clearly AMD's branch prediction is better, because this one number is bigger. "Once more I am disappointed by Intel"

While it could very well be true that the AMD branch predictor is straight-up better, the data they provided is insufficient for that conclusion.

vlovich123 10 hours ago | parent [-]

You may want to look up who Daniel Lemire is and the work he's done. What he's basically saying is "in the totality of things I've examined where Intel has come up short, this is another data point that is in line with their performance across the board". It's not "this one benchmark proves Intel sucks hurr hurr" - it's saying it's yet another data point supporting the perception that Intel is struggling against the competition.