http://www.digitimes.com/mobos/a20071224PD200.html

So AMD claims they need over 6 months just to get rid of the TLB bug?! Or is the real reason poor yields at the higher clockspeeds?

why the hell are they delaying cut down quad cores?
i understand about the whole tlb effecting everything but it hasent stoped them from shipping low end desktop quads,
are they trying to sink the company?
liek OMG WTF.

Better get a Phenom now ... might be a collector's piece in years to follow.

Or maybe they are now getting really cautious so as not to trip up ?

Or maybe they have respun some magic and about to unleash an astonishing miracle on us all?

Or maybe Digi are just spinning some Fuad?

Glad Toms and Anand are just looking at the facts ... Iv'e had enough of the dribble from tertiary sources that seem to know diddly.

One thing is for sure and that is a 280mm or so core must be difficult to make - only Intel's Pressler and Gallatin were this big (apart from AMD's first real core way back) and they generated heaps of heat. Mind you they were also 65nm and faster too.

AMD stock two for a penny... anyone?

I don't think I've ever seen a company shoot itself in the foot so many times over and over again as AMD has done in the past 18 months. It's a wonder that theyre still alive. If it was any other company Wall St would have beat it into the ground, but for some unknown reason its an investors darlin! It's going to be really fun to see Phinally Phixed Phreakinphenom debut just ahead of Nehalem, an arch that is going to make it look like a 486!

And the fastest 486 was an AMD from memory.

5x86133 ... clocked to 160.

Kept up with my mate's Pentium 75 / 90 ... his cost 6 times the price of my build.

386DX - 40 (AMD again)

P4 ... hello welcome to Thunderbird ... then A64 ... then X2.

Intel has only just gotten well and truly in front with the Core2 design ... built largely on the discarded P3 and mobile architectures with some brilliant cache re-engineering from those Israeli geniuses.

AMD hasn't got an answer yet ... a tricked up K8 just doesn't cut it ... even when 4 cores are snug as a rug on the same die.

They need to spend more money on R&D ... bit hard when your broke tho.

Intel will probably go back to sleep otherwise ... you will see speed grades in 10Mhz steps unless AMD survives.

This may be due to a lack of in-depth CPU manufacturing knowledge, but I don't understand why the issue would only affect chips at certain clock speeds.

Deleted

I agree with others that this is probably yeild issues, not TLB issues. AMD has shown that faulty chips won't stop them from shipping Phenoms to the consumers.

guys, a friend of mine in Idaho decided to buy a house at zero down. $350K when he was making barely $30K a year. Some criminal moron let him have the loan and a couple of years later he became just another statistic in the subprime mess. Now he's sleping on his moms sofa. Why the story? It's the same thing with AMD. Who the PhreakinPhenomHell ever told Hector the lunatic to go buy a company (ATI) that he couldnt afford? Fusion my a$$. You don't take on a $350K mortgage when you can barely pay your rent and you dont go swallow up a $6B company when you know your goin down the craper and your market cap will equal your buy in a year. And yes they did know it because there is no way that they were bilsfuly ignorant even back then that their roadmap was scifi. Yeild, TLb, whatever it was or is, I'm sure that there are plenty of PhDs at AMD who could have told Hector that what he wanted to do was unaffordable and technically too difficult. So instead of hunkerin down and concentrating on a strong mid and low market, they had to go out and blow up their own company by creating the utlimate chip... 40% clock on clock on Clovertown... blablabla. hector did this already in his last job so hes just doing it again. Corporate America is the only place where the more incompetent you are the more money you make and the better jobs you get.

Sure, AMD was a great copany and they made great chips. So what? Change one letter and you have AMC. They made some great cars too... the AMX was a killer musclecar. Then they started making Gremilins, Matadors and Pacers. Then they died. They desreved to die. Because they sucked. You can't live on your laurles of what you made that was great a decade ago. What's on the market now? Nuthin? Then you suck.

At least AMC didnt ship out Pacers with engines that were known to be faulty. AMD is! where the hell is the class action suit? If some manufacturer puts out a widget that doesn't work as advertised they are going to get crucefied. Why does the same not apply to AMD? Is ti because Wall St realizes that if they croak then Intel gets a monopoly and all hell breaks loose? There have got to be mroe reasons than that! Wall st is a free market system. They love to see the good rise and the bad fall. Why would they be subsedezing acknoledged bad managment? It's a mystery to me!

All I know is that Phenoms suck, and until B3 comes out as late as six months from now, they will continue to suck and if anyone asks me if they shoudl buy one I'll continue to tell them that they should have their PhreakinPhenom heads examined.

Responding to Reynod:

I agree with you. Everybody, look at the size of Intel's R&D departments around the world and then compare that to the few noobs AMD has running around in their little lab. Intel is simply a giant compared to AMD in terms of budget and manpower. They have R&D departments all over working on multiple projects at one time. AMD is a fraction of their size and was, and still is, the underdog. People don't understand that AMD was never an equal to Intel in terms of their ability to create new designs. I'll blame AMD fully for not growing more than they did in the few years they were on top. I guess Hector thought they could skate on by with their small company staying the small (relative to Intel's size).

AMD's only hope is that Intel will "go to sleep" and the giant is already showing signs of that. They could have already released a slew of new CPU's but they recently decided to wait a little. Their reasoning was that they did not have ANY competition in AMD. Intel has proven again that they won't come out with new products in a timely manner without competition. You Intel guys better pray AMD survives just to make sure Nehalem makes it out LOL...

I agree with you wingless. Why should Intel spend billinos on R&D for 32 and 22 nm chips when they can just keep selling the stuff they have now? Competition is the basis for inovation! But that does not mean that AMD has a licence to be incompetent jerks. The way it is now, the monopoly has already started. AMD simply does not exist in the higher end market. At all. Just because a company is there and fills up space doesnt automaticaly make it competition. they have to compete. AMD is not competing right now. Just cruising for a buyout so that their fat cat board can all get golden parashutes and go out and ruin some other poor unsuspecting company. Damn, what a scam. Wish I could do that. Unfrotunately i actually have to work for a living!

Hell yeah! I could eat Baconators for a week with all of that cash

Yeah you got a point. Only sue people who got the cash!

I think you mean to say this is the furthest AMD has ever been behind. Usually they release a matching generation architecture with superior performance than the 1- or 2-year-old Intel counterpart, and they can do this cheaply because of Moore's law. But they rarely released a top performer at the time. The 5x86-133 may have beaten P5-75, but in that same year Intel was selling P5-120s and P5-133s, also at stock. Why do you think the 5x86 cost so little?

To be fair, it's barely been 1.5 years since Conroe came out. AMD still has 6 months. But they appear strangely complacent. Perhaps they're content with more cores at lower clocks? I just feel that's the wrong approach, as cores don't scale perfectly, and people who buy so many cores also tend to want high per-core performance.

Digitimes inst a tertiary source, and known to be fairly reliable. Their reports, along with HKPEC, are often the primary reference for many of both THGD and Anands reports, not to mention the Inq and many other sites.