that's quite impressive half the power consumption at the same clock speed!! something tells me that 4GHz oc isn't that impossible after all.
Printable View
that's quite impressive half the power consumption at the same clock speed!! something tells me that 4GHz oc isn't that impossible after all.
Way to go,edit his own post to make him look like a homosexual...:rolleyes:Quote:
Originally Posted by gallag
Dude do you have limits while you're on the public forum?:confused:
:ROTF:Do they have a sense of humor where you live?You got to admit, Its a pretty funny sentience even without editing.Quote:
Originally Posted by informal
I would be happy enough with just the bug free fully optimized Deneb tomorrow, I would imagine that on the hole it would be easier lol (get it)Quote:
Originally Posted by Rammsteiner
what are you trying to say here?Quote:
Originally Posted by Karolis
I do get it.But it's distasteful.Get it?Quote:
Originally Posted by gallag
No actually, sigh, I fear you did not get it. I did not write (get it) as in "you get what I am saying mofo", I meant did you get the joke.Quote:
Originally Posted by informal
Notice the spelling mistake, On the hole should be whole. Did you never hear that old chestnut "On the hole, I find cheep toilet roll to be quite harsh" No? ah forget it.
Well some folks consider ass jokes to be distasteful.
Do you think i'm stupid or what?I saw you smileys but it's obviously distasteful,at least to me.I have pretty good sense of humor,thank you.
I don't really see it as being offensive. It's just a joke, and a pretty light-hearted one at that. We're adults here, aren't we?
Maybe I was trying to say that Devil's Prophet made a mistake... :confused:Quote:
Originally Posted by ryboto
It's over ninethousand!!!!Quote:
Originally Posted by ryboto
No, but srsly, the graph states that the 45nm Phenom uses 57,... ish watts, while the 65nm (9650 supposedly) uses more than 100 watts. I know the numbers seem off by 10 or so, and that is what he tries to correct. He says: 45nm power consumption is 95W minus the difference in the graphs. Or isn't he?
What now? The graph states ~57 W for the 45nm Phenom, and ~104W for the 65nm Phenom. I cannot believe the graph states anything other than sole cpu power consumption at full load. Those numbers are only off by some 10%, so PSU inefficiency can be ruled out I guess. So where am I wrong?Quote:
Originally Posted by Karolis
They probably measured the current going through the 4-pin, or 8-pin power connector. Since it's just a 12v line, if you know the current, you know the power. Problem is, they'd have to be measuring the power required for the CPU+VRM. So, in reality, the CPU will draw less than the reported numbers. Say the VRM is like 75% efficient, then each processor is really only using 42.7W and 78W respectively. Which sounds more reasonable for the 95W 9650.Quote:
Originally Posted by Devil's Prophet
OK, sorry... My bad... :) I know it sounds stupid, but I thoudt that was system's power consumption (I must have drunk something good yesterday :shakes:). But it would be great if a system with a 45nm Phenom and a 4850 consumed only 57 watts.Quote:
Originally Posted by Devil's Prophet
Yeah, that seems plausible.Quote:
Originally Posted by ryboto
Also: what percentage of the TDP can be reached when fully stressing an AMD Phenom? I've been hearing stuff about TDP, that it's not the average full stress power usage, but the utter limit the cpu can reach while doing the heaviest stressing session ever possible.
And what is the average efficiency of the VRM's? Does the number of phases have anything to do with that? Etc.
Intel and AMD define TDP differently, but yes--TDP is generally an indicator of worst-case scenario. Average power consumption is going to be less. Sometimes it will be much less.
I have no idea what the efficiency of the VRMs could be, I just threw out that 75% number randomly. I imagine higher phase circuitry might be more efficient, but I think the main purpose is for tighter power regulation.Quote:
Originally Posted by Devil's Prophet
MOSFET VRMs are generally very efficient. Just look at the cooling on the mosfets for example. On many boards that can do 150W+, there isn't any cooling assembly at all. It wouldn't take but a watt or two being dumped as heat by each mosfet on such a board to go POP.