According to our tests, this CPU is based on highly anticipated B3 stepping technology. And, it's meant to be error-free, allow higher clock speeds and still consume less power than AMD's Phenom 9300 and 9400 models.

http://www.tomshardware.com/2008/0 [...] index.html

hold your horses. As the article mentions, this is a "First Look".
Wait a couple o'weeks until a full review and then we'll see.

Can't wait for B3!
If there is a 9950, count me in!

#1) It fixes the Errata.
#2) It does not OC any more.
#3) It does not perform any faster clock for clock.
#4) It does not use any more power.

Gotta what to the Fall for the 45nm versions for those hopes.

So from the article I am guessing its just a Phenom 9700 just TLB errata free?

I just can't remember

did AMD claim it would give added performance or only people on forum wished it???

How do we know it doesn't OC any more or perform faster?
Certainly not from this article that's basicly "You know AMD had errata, they're gone in this new revision and it fits into a MSI K9A2 Platinium without a bios update."

We still don't know if AMD have done anything else to the Phenom in Revision 3, than getting rid of the errata. They might just have tweaked it so it performs better than the second revision in one or more ways, hell it might even be a decent overclocker. We just can't tell from the article.

However it's easter so I forgive Toms Hardware for this severely lacking article that's hardly interesting for anyone but owners of a MSI K9A2 Platinium mottherboard.

Did you read the whole thing? They state they will have a full batch of test in a review soon. Anand had the same thing yesterday so its not just THG. Next few days will be interesting.

Finally ...

I wonder, if we all think together that the B3 phenom will perform and OC much better, if it will really happen. Put those possitive brain waves into good use...

Well we don't know if it OC's further and we don't know for sure if it uses the same /more/less power....put we should expect the same.

What we do know is that the TLB "bug" which has only showed up in a -lab- so far is fixed.
To be honest this "bug" has never been seen outside of a -lab-.....so good call for AMD fixing the -bug- anyhow after first recalling past shipments.

Perhaps near end of the year they will ship a 3Ghz product.
No matter the maker in real life after you hit 3Ghz only a benchmark can tell if a game runs any better...and almost all software is still P4 rated.

I just want all the damn prices to go down on all the chips period. Intel needs some competion hopefully this brings down the chip costs all around.

I am sure it will.
Right now a game running with a AMD 6000+ @ $111 USD runs just as well as far as the player can tell vrs a top of the line Intel quad overclocked @ $1,100 USD.

Benchmarks can see a gain...but the game player or human eye can't.
Game and office software is still using the P4 rateings.

Thing about the OC'ing is we don't know that the TLB Bug was causing that. It might just be the architecture or the process(65nm SOI).

And for some games such as Crysis, even after 3GHz it helps. And a lot of games after that will probably be the same. But of course thats just one game and you are right that after 3GHz on most games only a benchmark will tell. I ran the CS:Source stress test twice and got 140-160FPS average. But to me it just looked normal until I saw the results.