zenmaster :
Well we know a little from this article.
We also know a little from AnandTech's Article.
If you read the article, Tom's did do some testing.
They found that the 2.4 used a little more power than the 2.3 version.
That would be due to the fact that the voltage is set higher in that system than stock voltage. My guess would be they didn't clear out any previous OC settings. But the drop in part would be to make sure the TLB fix wasn't going to effect the b3 as well, which Anand pointed out, the option completely disappears in bios with a b3 phenom.
zenmaster :
Clearly with such a small bump in speed if there was any power savings to be found the results should have been the same or less power used.
More power is being used due to the higher Voltage.
zenmaster :
Tom's also stated that the 2.4 showed an increase n performance that would be expected in a bump from 2.3Ghz.
AnandTech's article which was more indepth revealed the same things.
Actually the Anandtech review used the ES clocked at 2.3 to compare it to the 9600be clocked at 2.3, and showed the b3 performing slightly better on the test they used, likely due to the fix for TLB errata 254 possibly being fixed in b3 as well.
zenmaster :
The only performance gain you can claim is that you don't need to have the TLB Errata fix enabled in BIOS.
Note: Most home users did not enable this since it mostly effected Virtualization stuff. So if you want to claim B3 is faster since that BIOS patch is not required, I will agree but most sites reviewed the Phenom w/o than enabled for most tests and provided seperate results for when the patch was in place.
This is true though, granted a lot of the bios's that allow disabling the TLB fix don't implement it properly. A lot of the ones for the k9a2 plat that allow you to disable it only disable on the first core. Thats great for single threaded apps such as most games, but anything multithreaded you can tell, like winrar, the few multithreaded games, and encoding. Performance beta P.0j on the k9a2 plat is the first one I've seen that implements the disable option properly. But, a lot of the things with the phenom seem to be bios maturity related when it comes to stability. I've noticed that as the bios on the k9a2 plat matures, I've required less and less voltage to OC to the same point. For example under bios 1.1b3 I was running 2.6/2.7 at 1.262vid (1.248v actual) without C&Q enabled for stability. And now with bios P0j I can run the same speeds at 1.250v VID (1.240v actual) with C&Q enabled and completely stable. Stock voltage for a 9600BE is 1.232v actual.
zenmaster :
The Exciting news for AMD with the B3 is that they will be able to start shipping Server CPUs now. The B3 is of little importance to the Deskopt Market since at most it will usher in a new 2.4Ghz vs 2.3Ghz Phenom. Some point down the road a 2.5Ghz Version may also come out on the B3 Stepping.
They're supposed to be releasing the 9150-9750 and 9850BE within the next month or so. The 9850BE is a 2.5ghz B3 revision. No idea when they 9950 will be coming out, Phenom FX 82 is supposed to be coming at 2.6ghz but no idea when on that for sure either. Sorta makes me wish I'd waited for B3, but then again I haven't exactly had bad luck with my b2 9600BE.
zenmaster :
However, AMD has stated that faster Phenoms will not ship until 45nm comes out.
I wish the B3 had done more as it would help curb Intel prices.
I don't know, I think they'll hit 2.6 on 65nm, but I agree they probably won't ramp past that till 45nm as far as actual retail release speeds go. Though the BE's and FX's may be able to push past that with OCing. I can push beyond 2.6 right now, but mines an exception and not the norm. Granted that could be due to any number of things, cooling, airflow, or whatnot.