Phenom power consumption

[thilanliyan]

http://www.itocp.com/thread-11603-1-1.html

 

[Shortass]

Translated: http://translate.google.com/tr...ie=UTF8&sl=zh-CN&tl=en

 

[Lonyo]

That's pretty damned impressive.
7w drop in power at idle and a 24w drop at load just for the processor.

 

[PlasmaBomb]

Hmmm... the pins are different. Very interesting. They are also both running at 1.224 V so the power saving isn't due to dropping the volts...

 

[BLaber]

Originally posted by: thilan29
http://www.itocp.com/thread-11603-1-1.html

Thanks for posting this dude.

Impressive.With these early power saving number can we say that phenom will at least be competitive with Penryn in power saving ??? Any one , Any idea ?

Really looking good so far ==> 45nm X4 Phenom

 

[Extelleron]

Originally posted by: PlasmaBomb
Hmmm... the pins are different. Very interesting. They are also both running at 1.224 V so the power saving isn't due to dropping the volts...

Yeah, the really impressive thing here is that they are at the same voltage. If we saw a production 2.3GHz Deneb, it would surely have lower voltage, so the power savings would be even greater. And remember that this is an RB-C0 Phenom.... the first revision out of the fab shown to the press back in March.

 

[nerp]

Between their 45W cpus, good integrated chipsets and now this. . . AMD seems to be on the right track.

 

[SlowSpyder]

In my opinion this is what AMD needs more then absolute performance. If they can get their power consumption down, and of course clock speed increases won't hurt, I think they'll be on the right track.

 

[Idontcare]

Originally posted by: Lonyo
That's pretty damned impressive.
7w drop in power at idle and a 24w drop at load just for the processor.

What bothers me about taking these power consumption numbers to mean anything is that the Deneb of discussion was a pre-production sample and performance was not determined or disclosed.

Without performance (benching) at full-load being determined we can't really know for sure if the pre-production CPU had some features disabled by micro-code or fuse-blow at test just to generate some semi-functional units.

Of course this does not have to be the case, this pre-production Deneb could very well be fully functional and when fully loaded it is powering up just as many (if not even more) functional logic units as the Phenom CPU.

But the fact that the reviewer had both samples, ran the idle/load tests to generate power, but did not post benchmark data makes me very very suspicious that the load-power for the 45nm CPU is actually meaningful.

 

[Extelleron]

Originally posted by: Idontcare

Originally posted by: Lonyo
That's pretty damned impressive.
7w drop in power at idle and a 24w drop at load just for the processor.

What bothers me about taking these power consumption numbers to mean anything is that the Deneb of discussion was a pre-production sample and performance was not determined or disclosed.

Without performance (benching) at full-load being determined we can't really know for sure if the pre-production CPU had some features disabled by micro-code or fuse-blow at test just to generate some semi-functional units.

Of course this does not have to be the case, this pre-production Deneb could very well be fully functional and when fully loaded it is powering up just as many (if not even more) functional logic units as the Phenom CPU.

But the fact that the reviewer had both samples, ran the idle/load tests to generate power, but did not post benchmark data makes me very very suspicious that the load-power for the 45nm CPU is actually meaningful.

They already showed SuperPi benchmarks & overclocking : http://www.itocp.com/thread-11103-1-1.html

Performance is higher than Agena per clock by a decent margin, in SuperPi at least.

Of course it is an early sample and you can't judge final Denebs by what you see here. But I'm pretty sure that the results when Deneb CPUs launch will be even better.

 

[Accord99]

Originally posted by: BLaber

Originally posted by: thilan29
http://www.itocp.com/thread-11603-1-1.html

Thanks for posting this dude.

Impressive.With these early power saving number can we say that phenom will at least be competitive with Penryn in power saving ??? Any one , Any idea ?

That seems doubtful considering the QX9650 uses around 65W of power at full load. It looks like Deneb would finally be competitive performance/watt wise with Kentsfield.

 

[nerp]

I love google translate.

Heard that the use of the new SB750 South Bridge can be a substantial increase in non-agricultural overclocking ability! ! ! ! !
http://news.mydrivers.com/1/111/111820.htm http://news.mydrivers.com/1/111/111820.htm

non-agricultural must mean "not watercooled" haha.

 

[Shortass]

Originally posted by: nerp
I love google translate.

Heard that the use of the new SB750 South Bridge can be a substantial increase in non-agricultural overclocking ability! ! ! ! !
http://news.mydrivers.com/1/111/111820.htm http://news.mydrivers.com/1/111/111820.htm

non-agricultural must mean "not watercooled" haha.

Hah, I saw that but was wildly confused. That's a great way to phrase it!