Phenom Exposed! Shipping with flaky 3rd cores.

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar

They have. It's called Athlon X2 :)
 
I Just tested this "faulty core 2" theory with a phenom system im working on. Unfortunately ALL of the cores have the same ceiling of 2.55 ghz. (jacked the V up to 1.55) It could just be the biostar board I have in it, but from what i can tell at least on this BE, there is no one faulty core.
 
Rage,

Thanks for posting this information. The poster who accused you of not doing your research is sorely mistaken.

Perhaps you are correct and a tri-core will be a K10 with cores 0,1, and 3 turned on. It makes you wonder when products suddenly appear on roadmaps that weren't there originally.

Please keep us updated.
 


What version of Windows are you running on? Also, does anybody have issues on any other version of Windows or another OS like Linux, Solaris, or BSD? It could very well be an issue with the specific version of the Windows CnQ driver or that specific version of the Windows kernel not liking the Phenom's core synchronization or something and not actually be a hardware issue at all. Popping in a Linux/BSD live CD into a computer and seeing if you get problem is a very easy way to tell if it's a software problem because there isn't ANYTHING similar between the OS, software or drivers in those OSes and Windows. If you decide to do this, a very critical error will cause a kernel oops (you'll see a black screen with some error messages) or a hard lockup (rare.) Non-critical errors will show up by using dmesg or looking in /var/log/messages and they won't affect the functioning of the system much if any as they would in Windows.

You always have to try and isolate variables in a problem to make sure it's really what you think it is. Otherwise you can jump to wrong conclusions and not really get anywhere in fixing the problem.
 




dude!!!!
 
Also... seems u guys need KRAZY volts to get anything out of these chips.
1.5, 1.6 Thats INSANE

My $h!tty B3 Q6600 @ 2.7Ghz drops to ~1.05v under load and is 100% stable
 
Tis sad indeed. But as you know that is life in the big city. Consumers get and are getting screwed on a daily basis due to crappy products. Just look at the products that are recalled. Unfortunately it is the customer that usually gets the $hit end of the stick............Be an informed consumer. Intell had some problems in the past and have recalled the cpu's that were the culprits you would think that if the problem was that wide spread AMD would do the same.

http://www.recalls.gov/
 
this situation is bad

its too bad that amd tried to jump to the true quad in the media first and did not do their homework

we can only hope amd turns it around

me? i can not sell these system - its sad
 


I'm not too sure about the crazy volts thing. The phenom ive got (well its not mine in 2 days, so ask questions now while i can screw with it) runs 2.5 at 1.25 V... the extra .3 of voltage only affords it 50mhz of headroom (go figure)
 


I think that's an unfair comment. Sure, AMD has problems and no one should buy a B2 stepping Phenom, but we can't say, without evidence, that AMD is releasing them knowingly. It could be a quality control issue at their fab, an issue with 65nm process, or a bad design to begin with. Let's look at the possibilities.

Quality control issue:

Some Phenom's that should have been marketed as triple core, with the failed core disabled, get out into the quad core market. That's a QA issue that AMD needs to address. They should replace the bad Phenom's out there or provide refunds.

An issue with the 65nm process:

AMD is gearing up to bring out 45nm early, and I think that's due to issues beyond the errata with the 65nm process. As is, I won't buy even a B3 Phenom, but will wait for 45nm. There are plenty of viable X2 CPU's for this year's upgrades. AMD made a mistake in going native quad core at 65nm as there seem to be yield issues, and there are thermal issues with Phenoms.

Bad design:

I don't think this is an issue. Native quad core has the potential to be better than Intel's initial response of two dual cores in one package. Each individual Phenom core is 17% to 25% faster than an Athlon X2 core. The design is not as innovative a "next gen" as C2D was compared to Netburst, but it's sound.

AMD did screw up with the errata, but even Intel CPU's have errata. What happened with these cores appears to me to be quality assurance issues; we already know that AMD would market triple cores and there's no reason to disable a good core to provide a CPU for a market segment that might not even exist. Triple core implies, from the get go, that there are yield issues with some Phenom's experiencing a single bad core. Perhaps it's the same core on all CPU's?

Still, I do not think that AMD willingly, and knowingly, is selling bad CPU's. They screwed up, that's all. B2 Phenom is their Netburst, but they aren't like the Chinese knowingly using lead paint in children's toys. These bad core Phenom's should have had the core disabled and been marketed as triple cores.
 
It sure is easy to tell the Intel inside Idiot outside croud in this thread.
It is not a masking problem, nor a design problem, or it would show up in every chip. It is not a process problem, or it would have shown up in 65 nm X2s. It is not a QC problem, or it would show up in all other product.
It is most likely a motherboard problem. It can probably be fixed with a bios update.
On the other hand, the stupidity problem that a lot of posters in this thread have, will never go away.
 


can you elaborate, or would that request just offend u even more?
 


What a load of bilge, thats like saying "I know a secret, you would like to hear it, but im not telling"
 


LOL so much for a monolithic design being the "better" way to go, boy did that get old quick.

I still believe AMD has been having all sorts of issues with 65nm, considering just about all high end K8's are 90nm parts.

As for the Core2 issue, perhaps its cross talk with another part of the processor, or because of the L3 dishing out for 4 cores...

 


<rant>
Sure is also easy to tell the guys who *work* in semiconductor process design and manufacturing and understand these issues from those that don't.

Tell me, how would a motherboard BIOS issue create a flakey indivual core problem?
</rant>

It certainly can be a mask+process issue. The thing of it is that, even with a defect, it might not appear the same on every wafer. Believe it or not, the biggest issue facing those CPU manufacturers is variability. And the variability between lots, wafers, and even withing a wafer itself can be quite large. I'm more familiar with variabilty that can come from the patterning process (scanner dose, defocus, topography ("flatness" of the underlying features), scanner hotspots, temperature control, developer puddle, random etch residuals....)

Especially when given a portion of a design which is particularly prone to failure under process variability.

And it's not true that, just because you don't see it in the Athlons doesn't mean it's not a overall process problem. Certain design layout geometries are particulary poorly suited to handle the variabilities. No single design will explore all possible design space. AMD is at a particular disadvantage here becuase of the larger disconnect between design and process development.

I am leaning toward, not the design itself, but what it on the reticle itself that's the problem -- assuming the reports of the problem being dominant in core #2. And assuming the design itself is symmetrical. You see, even though the design of the four cores might be the same, it is NOT a guarantee that the data on the reticle will be. There are a number of data processing steps while taping out the reticle which subtly alter the design, and if the error is in there.

but hey, what do I know. I don't work for AMD and I don't know the details of the issue.
 
Well, I think it could be a BIOS issue or it Could be a Software Issue (Windows) that some have mentioned, if you ignore the fact that changing the CPU resolved the issue.

 


Definitely points to a design issue.
 
Well, from those 2 threads that were both started by the exact same person, I can gleen that most people are having problems on Windows Vista. Like I said I run Windows XP Pro sp3 and haven't had that issue. And since they're running on Windows Vista, they're using a hack/faulty work around to try and use AMD Overdrive.

And as far as Overdrive is concerned I won't touch it, since I've had nothing but trouble from it. It could also be a bios that doesn't have good control over the plane state settings though, I've seen a difference in stability while OCing or stock just in the k9a2 bios releases.
 



Must...defend...AMD...at...all...costs...
 

Blaming AMD's manufacturing problems on Vista? Cool :sol:
 


Yes, I'm blaming AMD's manufacturing problems, that apparently don't exist since they don't effect ALL phenom CPUs, on something other than the processor. And I'm not blaming it solely on Vista, I'm blaming it on people trying to use a hacked version of a program thats already craptastic to begin with, so they can run it on vista.

All I have to do is point at my sig as my proof. And I can even provide screenies. But no way in hell will I run AMD overdrive on my computer...
 


Same here. And I agree much respect.

While initially I have only heard the universal "TLB" issue repeated over and over with 99 percent of the repeaters having no clue what it means, It wasn't enough to sway my opinion in that the chip would be better with a new stepping. But this is enough of a legitimate case here to make me not go there. Even with the B3.

I am definitely going to an intel quad core next time around.
While I can understand selling a chip with erratum, as a lot of them have had in the past both intel and AMD, I cannot buy a chip with a bum core. That's Bull.

It also adds weight to the reason they are marketing tri-core chips too.

While I hope AMD sorts this out and gets back in the game, they will not do it on my dime while I am sitting here with a chip that is all effed up giving me BSODs. No way man.

I'm glad I went with the 5000+be for now, and I am sure I will be much more happy with a q6600 or a q9450.
Oh well.

 



I have also on occasion stood in defense of AMD in the past because I realize everyone screws up and deserves a second chance.

But this is completely unforgiveable. Erratum was enough and the discounted price made up for it...sort of.
But to sell processors knowing there were bad cores is a breach of my trust in AMD.

live and learn.

So....intel fanboy 😉 hehe
Should I go for a q6600 or do you think this q9450 is worth waiting for?

 



Must...buy...crappy...products...to...prevent...creating...monopoly! :pt1cable:

gotta love the amd fanbois though, for without you, I sure would've paid 500$ for this awesome q6600 (B3 no less!)

thank you.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom