TLB bug still present in B3 rev Phenoms

[gdogg]

Why do people keep saying crap like this?

Because of history, without competition , intel can charge (and has) whatever they want, for very little performance improvement.

Competition is what makes computers better and better, without competition innovation might not be invested it.

But on a serious note, I don't think the consumers would allow themselves to be taken advantage of this way again, if low end PC is costing $400 now, no one will buy a low end for $1000 in 2 years.

The prices of high end cpus, I don't think would go up, but low end cpus will.

[[XC] 2long4u]

It is the engineers playing stock games. They bought when the company was new. They made the 939, then sold their stock, then made K10. Now they are going to buy buy buy. Now they will fix the chip and make tons of money. lol

[K404]

It is the engineers playing stock games. They bought when the company was new. They made the 939, then sold their stock, then made K10. Now they are going to buy buy buy. Now they will fix the chip and make tons of money. lol

Thats the coolest "conspiracy theory" Iv heard for ages

TBH.....Its kinda tempting to join in!

[xVeinx]

While this is never good, I guess my question is how much it effects sales. What are the level of Opteron sales currently, and how much does the bug effect them really? Sure, they would likely get business from new processors, but is the level of Opteron sales enough to sustain them in the meantime? I'm not sure I'm ready to call AMD done yet. They have issues, but this isn't the first time they've had to struggle either...

[Cooper]

I can't really understand why are ppl give so much fus about this TLB errata thing. It is very uncommon to happen and as some resources mentioned can occur only in virtualization environment. Correct me if I'm wrong, but isn't virtualization only used to make many working places from fewer machines? And what tasks would be performed there? Office applications most probably. Now does any one gives a damn if word can accept 10k symbols per second or only 8k if patch is applied?

Though I still think that first respin of B3 was not concerned about TLB errata fix and there would be a proper one.

[Swarm dude]

As far as I am concerned, I didn't know that the errata doesn't occur always. But hey, I'd rather choose a CPU that has no bugs, at least we don't know if it (or them) has bugs. For example the K8 Athlons and current Intel's lineup. Why would I spend my money on merchandise with an factory error? Sure, I don't know will my CPU have that error, or if it has it, I don't know will I'll ever have problems.
But, that leaves a pretty bitter taste if you know what I mean and makes people avoid it
I don't know, maybe the sale results would be better if the current Phenom line up is cheaper. The 9500 has almost the same price as the Q6600, and the Q6600 overclocks better and is bug free. So people pick the Q6600 (or sth. else) instead...

[Cooper]

But hey, I'd rather choose a CPU that has no bugs, at least we don't know if it (or them) has bugs. For example the K8 Athlons and current Intel's lineup.

There are NO CPUs w/o erratas. Since 8086

[terrace215]

I can't really understand why are ppl give so much fus about this TLB errata thing. It is very uncommon to happen and as some resources mentioned can occur only in virtualization environment. Correct me if I'm wrong, but isn't virtualization only used to make many working places from fewer machines? And what tasks would be performed there? Office applications most probably. Now does any one gives a damn if word can accept 10k symbols per second or only 8k if patch is applied?

Though I still think that first respin of B3 was not concerned about TLB errata fix and there would be a proper one.

Maybe people fuss about it (from an investment point of view) because no Tier 1 OEM will ship K10 servers until AMD fixes it. They are selling the bugged Phenoms, but that is damaging the reputation of everyone involved, fair or not.

B3 was supposed to fix the TLB bug. It didn't. Say hello to Q3 for first OEM server K10 shipments. If AMD hasn't gone bankrupt first.

To answer someone else's question, no, the K8 DC Opterons do not cut it in the server space any longer. Intel has had QC parts out for over a year now. I'm guessing AMD's server share will fall to 5-7% by Q208, and that will just crush their ASPs and margins, and explode their losses.

[terrace215]

There are NO CPUs w/o erratas. Since 8086

But there are plenty of CPUs w/o errata so serious that OEMs will not ship them.

The "every CPU has errata" argument is silly. Of course they do. The question is, how bad is any given erratum, and the answer is determined by looking at OEM (read: CPU customer) behavior.

BTW, it's "erratum" (singular), "errata" (plural)

[Cooper]

B3 was supposed to fix the TLB bug. It didn't.

You say that like B3 already is in local PCWorld. This news is just a rumour until B3 will be actually released.

[mXpS]

You say that like B3 already is in local PCWorld. This news is just a rumour until B3 will be actually released.

Like you said its just a rumor as of now but we can all agree that this isn't the kind of news and/or rumors we want to hearing from AMD at this point in time. Let's hope the GPU development goes a lot smoother.

[terrace215]

You say that like B3 already is in local PCWorld. This news is just a rumour until B3 will be actually released.

With something this dire for the company, AMD would be denying it to the press if it were false. And note that the source says they got the information FROM AMD employees at CES:

google trans:

Here, in bulk, some information that we manage to steal employees from AMD.

We were able to confirm yesterday that the 9650 and 9550, "débugguées" version of the first Phenom (see our article) are also repelled the second quarter. AAMD has indeed received its first chips using the stepping B3 (early silicon), the bug is still not fixed. A new "spin" is needed, which could push the Phenom in May.

The manufacturer had published a roadmap in which the next generation of graphics cards were listed by the manufacturer in 2009 in the evolution of Spider platforms. Even under torture, our interlocutor refuses to confess if the R700 did emerge before the end of 2008.

Side integrated graphics, 700 series chipset AMD should see its life extended until 2009 pending the arrival of processors "Fusion" that incorporate graphics and heart-core processors traditional.

The manufacturing process of 45 nm AMD (who will not use high-k technology unlike that of Intel) is still scheduled for the second half of the year. Initial results were "encouraging". Hopefully ...

[Ace123]

Cooper, There are two types of computer enthusiasts.

Theres guys like me and you who I like to call "Christmas" users.
We simply wait patiently for cool new things to come out, and its like christmas for us YAY.

Then theres guys like those that posted those comments in this thread that I call "Horse race" users.
They sit, eyes glued open, watching every moment, every move, ever scrutinizing, and betting. They bet early and call their bet outloud and wait

These are the two types of enthusiasts.

[.Tret]

Because of history, without competition , intel can charge (and has) whatever they want, for very little performance improvement.

Competition is what makes computers better and better, without competition innovation might not be invested it.

But on a serious note, I don't think the consumers would allow themselves to be taken advantage of this way again, if low end PC is costing $400 now, no one will buy a low end for $1000 in 2 years.

The prices of high end cpus, I don't think would go up, but low end cpus will.

Because of what history? When in history has Intel charged $1000 for a low end cpu and $2000 for mainstream? I don't recall that happening; even in the early pentium pro days (the time when intels prices were highest).

In the pre-k8 days, Intel's prices weren't as high as people think they remember them to be. In the pentium 1 through pentium 3 days Intel always priced thier processors the same. $200-300 for the low-end and $800 or $900 for the high-end.

Even if Intel was a true monopoly, they could not sell cpu's at very high prices because not enough people would buy them to cover the cost to make the cpu's. The market is what determines prices. If AMD were to die right now I doubt Intel would raise their prices much at all. The problem we would face with Intel being a monopoly is lack of inovation not pricing.

That is why when someone says that Intel would price thier low-end processors at $1000 if they were a monopoly or when someone says that nehalem cpu's would cost $2000 if they were a monopoly, they are either stupid fanboys or are just plain ignorant. (too many people)

[LowRun]

With something this dire for the company, AMD would be denying it to the press if it were false. And note that the source says they got the information FROM AMD employees at CES:

google trans:

Here, in bulk, some information that we manage to steal employees from AMD.

We were able to confirm yesterday that the 9650 and 9550, "débugguées" version of the first Phenom (see our article) are also repelled the second quarter. AAMD has indeed received its first chips using the stepping B3 (early silicon), the bug is still not fixed. A new "spin" is needed, which could push the Phenom in May.

R700 is alright, thank you. That's what the AMD's graphic department's reps told us. The manufacturer had published a roadmap in which the next generation of graphics cards were listed by the manufacturer in 2009 in the evolution of Spider platforms. Even under torture, our interlocutor refuses to confess if the R700 did emerge before the end of 2008.

Side integrated graphics, 700 series chipset AMD should see its life extended until 2009 pending the arrival of processors "Fusion" that incorporate graphics and heart-core processors traditional.

The manufacturing process of 45 nm AMD (who will not use high-k technology unlike that of Intel) is still scheduled for the second half of the year. Initial results were "encouraging". Hopefully ...

I inserted and bolded the sentence that strangely got lost in your translation

[R101]

Why do people keep saying crap like this?

If you payed for Pentium 90 the ammount of $$ I did back in the day, you would know. When there's no competition, good hardware is expensive. Intel would be stupid not to raise prices if it could.

[Shintai]

I think people forgot how much a 286 costed. Or a 50MB HD back then.

Back then your value end entry PC costed 4000-5000$ and not 400-500$ or less. And back then money was more worth than they are today. So with todays money it would easily have costed 10000$.

Or perhaps what some of the early TVs costed.

People should elarn history, and not just what they wish. Price is lowering, why? Competition? Only a tiny part. But the main part is VOLUME.

With no competition Intels prices would not be much different than today. maybe 10-15% higher. And they would still innovate is almost the same pace.

If not, other technologies would catch up and make the PC into the gadget segment. And without big volume they would earn as much. It can sound hard for some people, but Intel earns more by selling CPUs at 200$ than they do at 2000$.

[Skyline GT-R]

Another reason why CPU were that expensive back then was because they were much more expensive to produce! A 65nm die in a 300mm wafer isn't that expensive, but a Pentium I with a 800nm proces technology just was much more expensive to make. You had alot less CPU's in a wafer then they do now.
Also about $1000 Extreme Editions, even the P4EE costed $1000 and those got crushed by AMD's FX cpu's, still intel managed to sell those. I don't think intel will raise their prices that much IF AMD dies. We can now buy complete computer for less then $500. If intel would increase CPU prices again and computer would cost atleast $1000 or more, then computer sales would take a huge dive!
Innovation on the other hand will go much slower as there is no compition.
That beeing said, I do hope AMD stays arround, but this sure isn't good news...

[Katzenschleuder]

I really wonder why this bug gets this huge amount of attention?
Current 45nm Intel CPUs suffer several bugs (which can't be fixed with microcode update) which may be even worse than this as you can read in the recent specification updates.
For example, Harpertown and Yorkfield has an equally serious TLB bug, some mixed load commands couse system locks, FSB anomalies with 4 layer mainboard PCBs, and so on...

This whole "Phenom TLB bug = the sky is falling" crap is simply hype spewed up by people who don't know what they are talking about.

[Nedjo]

Well , this certainly makes you have a great deal of trust in their engineers.

The B3 stepping supposed to fix B2 suffers from the same problem.Did they bothered to do a proper debug or some are sabotaging the company ?I'm pretty sure most EE aren't happy in Austin.

Btw , you get the beauty of this : Tier 1 OEMs won't touch the chip until this bug is fixed.This will do wonders for AMD's MS and credibility among Tier 1 vendors and enterprise customers.

I'm amazed how passionate are some people on this forum with every fart concerning the "news" about AMD's fall from glory!

Suddenly some unknown French site about multimedia is reliable source for ol' Intel boyz at XS! Pfff

BTW

http://www.x86watch.com/news/hp-phenom-desktop-200.html

[hollo]

How many times do you need to be told that the U.S. government, not to mention the EU is going to allow Intel to have a monopoly on the X86 license? It's just not going to happen. They'll be forced to renew the agreement with a probable suitor or come to some type of financial agreement.

if Intel had a monopoly on computers, then yea governments wouldn't allow it. but they'd only have a monopoly on x86. it'd be a long expensive legal process to decide whether that was enough of a monopoly to justify government intervention (eg splitting Intel up if AMD disappeared, or overruling the x86 contract so that if a company bought AMD they would get to keep the license). i don't think you can assume it would happen.

consider microsoft, for example. a near total monopoly on windows (like Intel would be a near total monopoly on x86 without AMD), but the government didn't have a good reason to break them up because there are other operating systems.

expecting the government to intervene if Intel was going to become a monopolist in x86 is just a bit too much like expecting the government to break up coca cola if pepsi went out of business (and of course the government wouldn't) for it to be a sure thing.

[Jimmer411]

IF the bug is still present, then AMD should just ditch the L3 and increase the L2 cache sizes in the next revision. If it's that easy of course.

I really doubt that this situation is as bad as "we" make it seem. There are countless companies that go thru the same thing, and many that still operate after chapter 11. Kmart is still around... There will always be some sucker investor to lend a hand.

[terrace215]

I inserted and bolded the sentence that strangely got lost in your translation

Sorry, it was unintentional. When you "copy" text from the google translation, it leaves the French sentences (which do not show on the page) spliced into the translation. So I had to hand-cut them out.

[SmaKKed]

I wouldnt say AMD is no more not yet, give it till the middle of the year and see what happens, they still have GPU sales which are doing pretty well and on a OEM level for mobile GPU (where i work) they are doing really well. But it is going to hurt them if it is true. Lets hope they get it sorted soon as i want to go back to amd for a while

[xlink]

I say all of XS pitches in and buys AMD when they hit $.02 a share