Retail Phenoms slower than review samples + more errata info

[Slobogob]

You all act like this is AMDs first chip and it means they can't make CPUs.

The truth is that AMD is acting like it is their first chip. Everything that could go wrong did go wrong. I hope they get their B3 stepping out soon.

 

[turpit]

You all act like this is AMDs first chip and it means they can't make CPUs.

I cant help but see the irony in that comment. It is in fact not AMDs first chip meaning AMD should have:
-Been able to deliver the product they claimed they were making
-Known enough not make claims they couldnt fulfill.
-Released a product which was..........ready for release, rather than ready for recall
-Known enough to know that they were subject to the same difficulties they gloated their competitors would experiance (which, ironically enough, their competitors did not experiance)
-Known enough to not commit the same acts they stood on their 'podium' pontificating their comptition was guilty of

One would expect an experianced manufacturer would have realized these simple things.

 

[dragonsprayer]

yes amd did not due their homework we all know
they bought ati to help solve their issues but the true quad has been harder then the cell it appears

 

[BaronMatrix]

So you mean I make posts about Intel? NOT!!! If they had a problem with 400M transistors, I'd say that's what happens. NONE OF THE REVIEWERS HAD THE PATCH APPLIED IN TAHOE AND NO ONE REPORTED CRASHES. quotemsg]

1) The systems were hand selected and controlled by AMD. They could have found the 5-10 most stable chips out of 10,000s they had built.

2) Even under these controlled settings, the reviewers were not allowed to attempt any Overclocks on the systems they were benchmarking. There was a single system they were allowed to OC, but they were not allowed to put any software on the system to test it. This indicates AMD feared placing any type of load on the systems. Not a good sign when you stage an event and can't even build a handful of powerful systems for folks to test.

3) The tested chips do not even match what is being shipped. They have adjusted the shipping chips to lower the NB speed, most likely to try and address the TLB issues but even that has not worked. They now need to shutdown parts of its functions. Once that is done, it will have little if any Clock for Clock performance advantage of the old K8 chips except for applications that can make use of all 4 cores.

Tahoe was the same as the Nehalem, Penryn and even Conroe first reviews. People were allowed to use the OverDrive application and some reviewers did do a test or two they brought on flash drives.

But if they did mean to defraud then someone should be punished. The fact remains that everyone sat quietly up until C2D like little sheep. I guess AMD would have to making $5B a quarter for you all to give it a rest. Well, with the increased share and XMas they may break $2B for Q4.

And since we insist, I can say that at least there are production problems unlike NetBurst which just sucked. And hey that so-called Hi-K is doing a great job at getting hotter than the 9650. Intel screamed from the top of their lungs how much better it was. And they do have enough money to throw at problems. I guess they just talk a good tick.

 

[BaronMatrix]

Tahoe was a freaking joke. AMD wasn't confident enough in their product to ship it out to reviewers and forced them to come to Tahoe and wouldn't let them raise voltages (read: OVERCLOCK).

We'll have to see Baron. I'm not convinced by one TechReport article, but I fear more will follow this week, if so, look to see AMD @ 8.50 a share

Maybe Hector is taking a page from the Steve Ballmer book. Microsoft was at $120 seven years ago and they make even more money now at barely $30.

AMD shipped Barcy to lots of reviewers. BTW, do you work for Intel and have to punish yourself for buying AMD?

 

[Scotteq]

Maybe Hector is taking a page from the Steve Ballmer book. Microsoft was at $120 seven years ago and they make even more money now at barely $30.

Baron: Please do us all a favor and educate yourself on what a Stock Split is. Then look up how many Microsoft has had in the last 7 years. Once you have accomplished those two things, you will understand how asinine your statement is.

 

[BaronMatrix]

I cant help but see the irony in that comment. It is in fact not AMDs first chip meaning AMD should have:
-Been able to deliver the product they claimed they were making

Why don't we talk about the effect of the losses? Can we not imagine that Intel knew full well that the price war would affect K10 by giving AMD less cushion to make less Brisbane and Turion.


-Known enough not make claims they couldnt fulfill.


Well, no matter what, a person can buy a functioning 9500 or 9600 from Newegg as of yesterday. I'd rather they pulled Opteron since SQL servers have mission-critical data

-Released a product which was..........ready for release, rather than ready for recall


If they had waited, it would be something else. Native quad core is more product differentiation than anything else. That's why Intel WILL do it. Then they have what AMD does. I have no doubt that AMD will quickly move to B3 now, which is said to not have the errata and to be a better stepping.

-Known enough to know that they were subject to the same difficulties they gloated their competitors would experiance (which, ironically enough, their competitors did not experiance)


Intel was the company without the guts to make it work. The losses contributed as I expected to this release.


-Known enough to not commit the same acts they stood on their 'podium' pontificating their comptition was guilty of


Which acts are those?


One would expect an experianced manufacturer would have realized these simple things.

 

[BaronMatrix]

Baron: Please do us all a favor and educate yourself on what a Stock Split is. Then look up how many Microsoft has had in the last 7 years. Once you have accomplished those two things, you will understand how asinine your statement is.

Well, as a former employee who watched the stock reach 120, I know that after splits the stock usually goes back up when Wall St is happy. Splits ALWAYS occur after the stock RISES consistently. MS' 5year is between 22 and 37. A stock that traded at 120 doesn't split when it's stuck at 25% of its former high. If MS did a 2 for 1, the stock would be at 16.

But then, what do I know?

 

[Grimmy]

You all act like this is AMDs first chip and it means they can't make CPUs.

Whelp, in a sense, it is their... First.... Native... Quad... CPU. 😗

 

[Scotteq]

Nice Tapdance there, BM.... Infering that MSFT would do a split now, when I was pointing out that they have previously split.

What you have done is take a PRE-split $120, claimed the current $30 is indicative of the stock's performance, and completely ignored the effect of the the splits which have occurred in the same timeframe. What this does is make stock that's more or less flat look like it's trading at 25%. It's bullsh*t, and you know it d*mned well. You are smarter than that, therefore I must conclude you are intentionally lying.

Besides - Microsoft *MAKES* money. AMD is currently not doing so.

But what do I know...

 

[NMDante]

Tahoe was the same as the Nehalem, Penryn and even Conroe first reviews. People were allowed to use the OverDrive application and some reviewers did do a test or two they brought on flash drives.

Nope. Tahoe was a controlled test environment for their soon to be released CPU.
Conroe results were done independently, and were not believed by many when the initial testing results were released. Same with Penryn (Yorksfields). There was no Intel controlled event.
Nehalem hasn't been reviewed by anyone, and was shown at IDF, which is an event by Intel for more than just benchmarking or testing.

But if they did mean to defraud then someone should be punished. The fact remains that everyone sat quietly up until C2D like little sheep. I guess AMD would have to making $5B a quarter for you all to give it a rest. Well, with the increased share and XMas they may break $2B for Q4.

$2B in Q4? They can only hope. And no, they don't need to make $5B a quarter, but even breaking even would be good for them.

And since we insist, I can say that at least there are production problems unlike NetBurst which just sucked. And hey that so-called Hi-K is doing a great job at getting hotter than the 9650. Intel screamed from the top of their lungs how much better it was. And they do have enough money to throw at problems. I guess they just talk a good tick.

You really don't read well, do you. Is the QX9770 released? No. It's still an ES CPU. If the QX9770 is still consuming more power when released, then you can laugh all you want, until then, that's just silly.

Netburst might have sucked, but it was still profitable. And I'd rather have a sucky working CPU, than production problems anyday. A process can only be tinkered so much, until a new process development is needed. Oh, and Intel isn't screaming about Hi-K, they don't need to. If it wasn't so great, why is IBM and AMD thinking about using it?

You have a knack of blaming everyone but the actual problem - AMD.

 

[TechnologyCoordinator]

Baron: Please do us all a favor and educate yourself on what a Stock Split is. Then look up how many Microsoft has had in the last 7 years. Once you have accomplished those two things, you will understand how asinine your statement is.

Thanks, you saved me the time and typing.

 

[TechnologyCoordinator]

Tahoe was the same as the Nehalem, Penryn and even Conroe first reviews. People were allowed to use the OverDrive application and some reviewers did do a test or two they brought on flash drives.

You are not correct. This was the freaking week before the product was recalled, opps, I mean launched, not recalled. Conroe was shipped to reviewers and they were allowed to use and abuse the chips on their own hardware and in their own labs.

AMD knew it had a dud and that is why the Tahoe event was staged instead of sending the chips to reviewers.

 

[Kamrooz]

Tahoe was the same as the Nehalem, Penryn and even Conroe first reviews. People were allowed to use the OverDrive application and some reviewers did do a test or two they brought on flash drives.

But if they did mean to defraud then someone should be punished. The fact remains that everyone sat quietly up until C2D like little sheep. I guess AMD would have to making $5B a quarter for you all to give it a rest. Well, with the increased share and XMas they may break $2B for Q4.

And since we insist, I can say that at least there are production problems unlike NetBurst which just sucked. And hey that so-called Hi-K is doing a great job at getting hotter than the 9650. Intel screamed from the top of their lungs how much better it was. And they do have enough money to throw at problems. I guess they just talk a good tick.

Interesting that you claim netburst sucked. It was a power house in terms of wattage, also in terms of heat. But hey..Taking my extremely hot 3.2 prescott to 4ghz was not a problem what so ever, at least it worked. It may have taken/produced a lot of wattage/heat....But atleast there weren't architectural issues that completely destroyed the value of their processor. Even compared to the Athlons at the time...The intel solutions were still better for video editing (single core era). Compared to the current barcelona's we have from AMD...Which were run through countless delays, only to be released faulty. It doesn't clock high, it has leekage issues, and MANY other issues which I feel bad to even mention. The K8 series were simply fantastic processors, I love the X2 series as well. But stating Netburst sucked, when AMD is having some massive technical issues with a retail Barcelona, is flat out childish. Netburst wasn't the best processor, we all know that. But atleast it worked without issues and clocked high to provide some competition.

I pray that a new stepping does help out bring Phenom to competition. But I can't understand how you can spew this garbage of netburst sucking and all this other garbage, when at the time they had processors retailing at the thousand dollar mark. Their processors worked, they could clock higher even though they demanded lots of power/heat, unlike the current Phenom which can't even clock above 2.3 ghz. Compared to now, the most expensive phenom is 260-300 dollars, which might I add is overpriced for it's performance.

Regarding High-K. if you want to be ignorant and ignore the fact that the QX6850, like the QX9650, is also rated at 130TDW, then that's your fault. The QX6700 is rated at 130W TDP, mean while the q6600 is listed as 95W TDP. These extreme edition processors always have a high TDP, unlike their desktop consumer based platforms which run cooler with the less voltage. Those 45nm desktop yorkfield processors have not been released to market yet, and will surface in january 2008. We have no idea what the power drawHonestly man, you really need to think about what your saying and look a little deeper before you just spout nonsense that makes you look even worse. We also know how AMD/Intel both rate power draw differently.

Hopefully AMD can fix these issues. I'm rooting for them, but I don't understand how you can not look at the facts and just admit it for once. The article approval concept and NDA which amd adopted before showing off barcelona, due to their inferior processor, was a complete joke. Not allowing reporters to post a legitimate review? Forcing them to submit it to AMD to approve the article before it can be posted? AMD unfortunately bit off more then they can chew. I will admit gladly that Intel has adopted decisions that AMD has taken in their design. Soon will have Nehalem with a IMC and their Quick Connect interface which is basically AMD's HT link. AMD always plans to innovate which I respect and approve of, but they bit more than they can chew, this time. They acquired ATI, they split their R&D staff and funding among vast amounts of projects, and blindly followed their Native Quad core ideology too far.

Regarding C2D, it gave enough of a reason for consumers to sit in front of it like little sheep. It was a fantastic processor architecture, and it still is. Everyone waited patiently for AMD to release Barcelona, no one was angry. But once a delay, after another multiple delays were laid on top of each other, people got angry. With good reason too, the product was hyped to no end, and failed to meet expectations. I won't even comment on their actual current retail products. They were forced to work on all these issues, but still couldn't get it done quick enough. With all the delays, AMD had no choice but to release a unfinished product if they wanted any chance to make profit. Even if that meant selling a incomplete product, which you cannot deny, is incomplete due do these issues. We have production samples for a reason, but the actual retail product which has been delayed for so long has become the production sample, which is outrageous. I do not doubt that IF AMD can solve these problems, and get their Barcelona architecture clocking high, that there competitiveness will rise. There faster clocked products will rise in price to compete with intels offerings. Will this happen soon?, we have no idea, but if it doesn't, AMD is in big trouble, and you have to at least admit that. It's not that difficult to look at all the facts, which there are many, and admit it. Penryn isn't a big issue, it's just the "Tick" in the "Tick Tock" program, which is the perfect way to go about the x86 market. Once the "Tock" (Nehalem) is released, AMD will either face the most difficult and worst scenario for them ever, or, if they can fix their issues before hand, they might make some money before Nehalem hits market. We have no idea how much of a dent this mass focus of R&D staff/funding on Barcelona has affected their entire road map. But it would be foolish to think it hasn't. If they don't fix these issues and make some money before Nehalem hits, there is a large chance AMD will go under without support from other companies/investors. That's a fact, and plainly visible, if you just sit down, look at the situation, and think rationally, it's easy to admit. There is nothing wrong in doing so, it doesn't mean you are insulting a company that you respect. Blind love, gets you no where.

It's all up to AMD, if they can solve the problems and get some higher clocked phenoms out. They might make it through with some profit. If not, Nehalem will end up sealing the coffin, they have around 12 months...Let's see how well they do.

 

[Kamrooz]

Bah..double post, my bad.

 

[spaztic7]

the_vorlon:

Your avatar is simply amazing!!!!

 

[spaztic7]

AMD stock has dropped 4.43% and dropping quickly. Although it does pick up for a second then drop back down.... almost like someone is buying stock....

 

[Shadow703793]

May be AMD should sell ATI and use the money to improve CPU's?

 

[Kamrooz]

They are in it for the long haul. Purchasing ATI was a smart move considering what AMD's aim is for. They just happened to purchase it at the worst possible time, around the release of Core which completely killed their earnings and has driven them into debt.

I just hope a simple stepping revision can fix their issues to attain higher clocks up to the 2.8/3 ghz mark. Cause if they can't raise some competition to pull in some funds before Nehalem is released, sad to say, it might be the end for AMD. They probably won't vanish, downsize tremendously and become a smaller company. But I don't think that will happen...Unless the situation is even worse than it seems....Which in that regard..is a very sad sight..

The day competition dies in the x86 market, the more we will pay for our products, and the slower the drive for innovation will become.

 

[spaztic7]

very true Kamrooz. But at the rate that AMD is going, we may not be able to see what they aimed for when purrchasing ATi. I for one, am really saddened by these turn of events.

 

[Kamrooz]

Same, the B3 stepping from what I've heard will have a March 2008 launch. This is a very scary ordeal. If AMD can't fix these issues and get Barcelona clocked high and performing at 100% capacity...They won't be able to turn some profit for more R&D...If that does indeed happen. We may see AMD fall out of the x86 race during mid/late 2009. I really hope this does not come to be. With the acquisition of ATI as well..it will be a very large consequence in the x86 and graphics card markets.

 

[d4NjvRzf]

I just hope a simple stepping revision can fix their issues to attain higher clocks up to the 2.8/3 ghz mark. Cause if they can't raise some competition to pull in some funds before Nehalem is released, sad to say, it might be the end for AMD. They probably won't vanish, downsize tremendously and become a smaller company...

...and, Turpit would say, become a wallmart pc supplier of choice! (jk)

 

[turpit]

turpit wrote :
--------------------------------------------------------------------------------
I cant help but see the irony in that comment. It is in fact not AMDs first chip meaning AMD should have:
-Been able to deliver the product they claimed they were making

Why don't we talk about the effect of the losses? Can we not imagine that Intel knew full well that the price war would affect K10 by giving AMD less cushion to make less Brisbane and Turion.

No, why dont we stick to the subject and talk about AMD:
Since it is the product AMD made which does not perform up to claims which were made by...AMD, not Intel.
Since it was AMDs R&D that created the product, not Intel's
Since it was AMD which released a product which was so far from being ready for release that one half of the product family has been recalled while the other half has been limited to minimize the impact of the problem affecting the design.

Baron, once again you use one of your typical tactics, attempting to divert attention from the problem by laying blame for the problem elsewhere.....and it is still a failing tactic. AMD designed, manufactured, promoted, and released the product, not Intel. That AMD failed to live up to their claims or provide a quality product is AMDs responsibility, not Intels.

turpit wrote :
--------------------------------------------------------------------------------
-Known enough not make claims they couldnt fulfill.


Well, no matter what, a person can buy a functioning 9500 or 9600 from Newegg as of yesterday. I'd rather they pulled Opteron since SQL servers have mission-critical data

And after 18 months of deciept and lies, failure to meet claims, a preemptive product halt, and an entire product line recall, exactly how many people do you expect will buy this flawed product? As many as purchased QFX perhaps?

turpit wrote :
--------------------------------------------------------------------------------
-Released a product which was..........ready for release, rather than ready for recall

If they had waited, it would be something else. Native quad core is more product differentiation than anything else. That's why Intel WILL do it. Then they have what AMD does. I have no doubt that AMD will quickly move to B3 now, which is said to not have the errata and to be a better stepping.

(A)
If they had waited, they would have saved money: They would not have produced retail silicon which had to be recalled...silicon which cannot be "fixed" but must go to the waste column
If they had waited, they would have saved face: Now, instead of rumours of difficulties, they through their own actions have established it as fact they they are having difficulities

(B)
Native quad core is 2 things to AMD:
1- Face saving.After all the pontificating they did on the 'failings' of MCM, they would have looked like buffoons for producing a MCM. But you know this.
2-The easier path. IMC does not lend itself to MCM. In order for AMD to go to an MCM, the easiest, fastest and most succesful route would have been to got to an off die memory controler...again, another blow to their "PR" ego/image. But again, you know this.

(C)
I find it oddly amusing that you would make the statment

Native quad core is more product differentiation than anything else. That's why Intel WILL do it.

When you know:
AMD reps have said they should have gone with an MCM prior to native
AMD has already said they are going to be producing an MCM.
Intel already has, and has had for over 12 months, a working quadcore.

(D)
Native quad at 90nm on SOI was so prohibitive as to be impossible. Native quad at 65nm on SOI was foolish.

turpit wrote :
--------------------------------------------------------------------------------
-Known enough to know that they were subject to the same difficulties they gloated their competitors would experiance (which, ironically enough, their competitors did not experiance)


Intel was the company without the guts to make it work. The losses contributed as I expected to this release.

"Guts to make it work"?
First, let me say Im sorry that I didnt realize this was a measure of the 2 companies penises rather than the value and quality of the products they produce. Since that is the approach you choose to take in response, allow me to provide you with a small dose of facts regarding your comment. AMD, in fact, did not make "it" work. Had they made "it" work, they would not now be recalling thier products, and in fact, the 2.6/2.8GHz products would have been released on schedule.

Now, that said, dont confuse foolhardiness with courage....they are not the same. It is more likely that AMD took the path they choose because of the negative press they would have recieved had they gone MCM, only because of their own PR regarding MCM.

Second, what, pray tell, does Intel have to do with AMD making anything "work"? Did Intel loan AMD one of its R&D teams? No? Did they let AMD borrow a research facility? No? Did they sell AMD the SOI technology? NO? Did Intel design K10 for AMD? No?

Please stop trying to divert attention away from AMDs difficulties by trying to blame Intel for them.

turpit wrote :
--------------------------------------------------------------------------------
-Known enough to not commit the same acts they stood on their 'podium' pontificating their comptition was guilty of


Which acts are those?

Oh please Baron, dont play dumb, you know quite well.

turpit wrote :
--------------------------------------------------------------------------------
One would expect an experianced manufacturer would have realized these simple things.