Re: Does leakage traditionally double with 15 degree C increase -- from 35C to 50C?
I've seen some sources claim that leakage can double with only 10C degrees of variation. For example:
interfacebus.com
Here is some additional data that suggests subthreshold leakage is strongly governed by temperature, going up exponentially as temperature exceeds ambient:
ieeexplore.ieee.org
Here is another link with a helpful equation:
powermanagementdesignline.com
At any rate, I only claim that the Intel and AMD specs for leakage are not comparable, due to being tested or extrapolated at different temperatures. I never claimed that the Intel specs are better, because independent testing has verified that AMD has a small but apparent advantage in idle power (certainly not as large as Pete would have you believe, and the funny part is that the specs verify it!).
Re: Yes. Can you point me to on the process specs? I'm really curious. Is it *JUST* personal opinion or do INTC and AMD have specs out there on this subject? You might want to avoid the use of LOL where it's personal opinion unless you want to get it back.
Maybe you should reread what I wrote. I described the Intel process of entering C1E, and the AMD process of entering C1E, and AMD has more steps involved in what they need to do. Yet Pete claimed without proof that the AMD way is shorter. I find that funny, but not quite as funny as you standing up for him when he is the one who is not honoring the burden of proof. All I asked was whether he really believes that AMD can do more power management operations than Intel in a shorter amount of time. That's a bold claim for something not explicitly stated in the specs.
Re: Whoops, Attack the persona. I know how to do that. Wanna see me try?
No, I'd rather you kept the discussion civil. You find it offensive that I am laughing out loud, and sarcastically pointing out Pete's foibles, but you don't seem to be offended that Pete repeatedly makes mistakes and passes them off as fact. By the way, several people have already corrected him that the right term is "Griffin" instead of "Griffith". I found it amusing that he continues to get it wrong, so I laughed out loud ("LOL"). Pete's posts are funny, what can I say...? He is constantly predicting that AMD will steamroll Intel, and these posts of his have been going on forever. Not even after all this time away from SI can I reason with his impervious headstrength. But I'm sure he appreciates you standing up for him. It gives him the motivation he needs to continue making false and unchallenged claims. |
