Intel Identifies Chipset Design Error, Implementing Solution

[qcmadness]

But the OEMs are probably doing accelerated stress testing with increased heat and voltage to simulate years of normal stress. If the controllers were degrading in a few weeks after normal use, I think Intel would have caught it with their normal tests

This is correct. They are using artificially temperature inflated testing using "oven testing", where air flow is cut off and extra component heat is heavily simulated.

There are reports from Asus / Gigabyte... etc of sata controller inconsistency.

[moshpit]

But two processor failures (Pentium bug, PentiumIII 1.13 Ghz recall) and now two chipset recalls (915G/P & 925X, now 6 series) from Intel don't disturb your faith. Not a fanboy at all

LOLOLOL!!!! I laugh at your mention of things I protested against and hated Intel for years over as proof I'm some kind of Intel fanoy. Until Conroe, I never even owned an Intel CPU, get a grip, son. Since Conroe, AMD has failed to give me a reason to keep buying them. That's not fanboyism, that's simply realism. Intel is faster over all and has been since Conroe. I now support AMD's GPU division when they give me a reason to, which seems to be every other generation or so.

Usually, it's the Fanboy who makes the first accusation of fanboyism, so get yer head out of AMD's butt already

Edit: Also, for the record, if Bulldozer beats Sandy Bridge overall when it launches, you can bet I'll have one. I'm just completely unconvinced by anything I've seen so far that will be the case.

[NEOAethyr]

If you guys don't end dis fanboy bs, da thread will end up getting 86'ed... ^^

[Sam_oslo]

It's the same story in this news section, everywhere, when AMD-fans doesn't like an idea, then they stamp with a fanboy. It starts with these kind of ignorance and soon develops to personal attack, offensive language and childish fights. At the end, the constructive arguments drowns in BS.

It's the same when people talk about AMD GPUs too. There are a few good old forum members, who are AMD fans, and cherish these kind of behavior, and then new people adopt it too, I've noticed.

[Tenknics]

Exactly what I thought and said.

This isn't as big of a deal as you were all making out to be. Haha, now that all the facts are out you realize this.


Btw package from newegg just arrived..

i7 2600k
asus p8p67 pro
gig 560 Ti OC
16gb Gskill
64gb c300
2TB WD black

[Touche]

LOLOLOL!!!! I laugh at your mention of things I protested against and hated Intel for years over as proof I'm some kind of Intel fanoy. Until Conroe, I never even owned an Intel CPU, get a grip, son.

Usually, it's the Fanboy who makes the first accusation of fanboyism, so get yer head out of AMD's butt already

I'm pretty gripped, nephew. And it would be you who made the first fanboy accusation in this topic.

On topic...

"Intel claims that only 5-percent of Sandy Bridge owners will experience the flaw. But speaking with manufacturers, we were told that numbers appear
to be far greater than that. Some companies told us that this is definitely a "panic-mode" situation for them."
http://www.tomshardware.com/news/san...a-3,12112.html

[initialised]

This is correct. They are using artificially temperature inflated testing using "oven testing", where air flow is cut off and extra component heat is heavily simulated.

This is a standard testing methodology, it's known as accelerated lifetime testing and has been used by the (mirco)electronics industry for years as a way of determining the products MTBF presumably their data suggests it is significantly less than the target lifetime of the product.

[initialised]

Exactly what I thought and said.

This isn't as big of a deal as you were all making out to be. Haha, now that all the facts are out you realize this.


Btw package from newegg just arrived..

i7 2600k
asus p8p67 pro
gig 560 Ti OC
16gb Gskill
64gb c300
2TB WD black

It's no biggie, ASUS aren't about to recall all 6 series products or anything...

/sarcasm

[NEOAethyr]

Well in a month or so you guys should beable to rma, which I recommend for everyone.
Not so you can run intel's assents in to the ground, but the truth is we don't need these kind of boards out on the market for people to buy.

The majority of people will probably never see this prob.
It really depends on how you use your pc, how much you stress the south bridge with your everyday stuff.

My recommendation of course is to not use the effected ports, especially for burners...
I don't know about you guys but my burner has to keep working.
(benq dw 1620 to my current burner in my sig, went through 3 crappy liteon's in between)

You could probably kill these controllers probably within days, using 24/7 non-stop random data block writes to the drives.

Which surprises me that intel didn't test them throughly.
They're a chip maker, not a drive maker.
I know that drive makers don't test there drives very well.
All they do is write FF's then 00's to the drive, if it can do that, it's good to them and they ship them out.
Some of us already know that is not a reliable method for testing.

Well at least now they will probably test them a little bit more before they mass produce them.
This was a big mistake on intel's part, so I'm sure there was something to be learned from it.

[Sam_oslo]

Yeah this has been an expensive mistake, and they will learn from it.

I guess Intel pushed the chipset-stuff a bit too hard. They got scratched, by working on 3 chipsets P67, H67 and Z68 all at the same time.

I haven't noticed any problem on my ASUS P8P67 Deluxe yet. I recommend Intel replaces our MBs with a Z68 . That would be a good juster and could make us early adopters happy.
EDIT:
Right now, they should announce everyday can chose to replace their MB with a Z68 when it gets available. It would put an end to frustrations and speculations, I guess.

[thematrixhazune]

I think it would be a good idea for some review sites to build a couple rigs and run them 24/7 or even just 8 hours a day running a batch list of apps. Each day or week test for bit errors ect. Record the results on a web cam for us to see. Would the increase of traffic offset the extra power cost to run?

[aytacaksel]

http://www.gigabyte.com/press-center...e.aspx?nid=984

Taipei, Taiwan – February 1, 2011 – GIGABYTE TECHNOLOGY Co., Ltd, a leading manufacturer of motherboards, graphics cards and other computing hardware solutions, today announced that GIGABYTE has been alerted by Intel about a 6 series chipset design error. GIGABYTE is working closely with Intel® to minimize the inconvenience to customers and retail stores with regard to related issues.

A recent statement from Intelindicates that they expect full 6 series chipset volume recovery in April, 2011. GIGABYTE will produce and deliver new motherboards with the updated Intel 6 series chipset when the new chipsets become available. GIGABYTE prides itself in having the highest service quality in the industry, and as such we will resolve this issue with minimal impact on our customers.

Additional information will be added to the official motherboard websiteand the GIGABYTE Tech Daily blogas the situation develops.

[initialised]

Just got a phonecall from my Asus tech rep, expecting sort of an official statement later today

Have you read it yet? Full recall, sounded a bit extreme if this is a minor problem that will only affect a minority of users. It's going a bit further than Gigabyte's statement.

So can you RMA the board even without explicit problems?

With ASUS you can, but you'll have to wait until the replacement products are ready.

[Sam_oslo]

Have you read it yet? Full recall, sounded a bit extreme if this is a minor problem that will only affect a minority of users. It's going a bit further than Gigabyte's statement.
...

This may be a bigger problem too, but not necessarily.

Intel may have chosen a total recall to keep their good reputation. That's actually a good quester. They have proven to be serious about keeping their products in top notch form.
Nothing suggests that it goes further than that GB-statement. They don't need to down-dramatize it, their total-recall act is dramatic enough. How could they take it further you think?

[RaZz!]

I think it would be a good idea for some review sites to build a couple rigs and run them 24/7 or even just 8 hours a day running a batch list of apps. Each day or week test for bit errors ect. Record the results on a web cam for us to see. Would the increase of traffic offset the extra power cost to run?

lol, 24/7 livestream! like big brother for hardware nerds.

"DUDE, I THINK IT JUST HAPPENED!" - "what? holy crap!" - "na, sorry, false alarm. it's just the screensaver, nvm!"

[Aberration]

Have you read it yet? Full recall, sounded a bit extreme if this is a minor problem that will only affect a minority of users. It's going a bit further than Gigabyte's statement.

With ASUS you can, but you'll have to wait until the replacement products are ready.

Well obviously LOL, whats the point of RMAing a product to get another effected product.

Is it just ASUS? Not sure why, I have never had to RMA anything, but I have always liked ASUS products, on my ASUS G50VT-B1 laptop... which I was going to replace with a new Sandy Bridge one in a few weeks when I get back to the States.... Damn I don't like waiting!

[Eastcoasthandle]

Hmm, I wonder if overclocking the BLK will be more likely once the new chipsets are released? If so, it probably won't be as much. Similar to what users got with the FSB of the Core 2's.

[Aberration]

I doubt any revision will effect overclocking much.

[thematrixhazune]

lol, 24/7 livestream! like big brother for hardware nerds.

"DUDE, I THINK IT JUST HAPPENED!" - "what? holy crap!" - "na, sorry, false alarm. it's just the screensaver, nvm!"

I wouldn't mind watching it! lol.
I remember years back Tom's did it with an AMD and INTEL rig.

Obviously some background crunching would be mandatory.

[SubZero.it]

@SubZero.it
That's probably because some people stress systems more then others.

Ok...but it should be rock stable in a such important part as the socket...and sata is not connected directly to the amount of stress...in fact they are retireing boards. Quality is quality. And in generation is missing I don't know if you are gonna trust in what I say, but bear in mind that I'm not a fanboy of any kind. You find some kind of objective here

[Sam_oslo]

I doubt any revision will effect overclocking much.

The chipset can actually play a big role for high-OCs, not much on 24/7 OC tho.
Many are hitting the x51-wall. This seams to be due to CPU-Batch#, but chipset/QPI could create such wall too. It remains to be seen how the "fixed" chipset will affect this tho.

[Tenknics]

connection error, myb

[Tenknics]

It's no biggie, ASUS aren't about to recall all 6 series products or anything...

/sarcasm

Where's the info on the Asus recall?


And again, I sound like a broken record, but if I'm not plugging into the 'affected' ports, than I will be fine. Nice icon btw, explains much.

P.S. the boxed sandy' B's are so tiny and cute


http://img214.imageshack.us/i/cutecpu.jpg

http://img593.imageshack.us/i/paulcomp.jpg

[initialised]

You're probably fine, just don't use the SATA-II ports and RMA the board once the replacements are ready.

Procedures and info will surface later today or tomorrow, basically they are recalling all boards in the channel and offering RMA to end users with affected products. I'd be very surprised if the full text, is not online somewhere.

PCPro The info here is slightly out of date, but I have received the statement they are due to release tomorrow and it is contrary to the penultimate paragraph.

[Hornet331]

The chipset can actually play a big role for high-OCs, not much on 24/7 OC tho.
Many are hitting the x51-wall. This seams to be due to CPU-Batch#, but chipset/QPI could create such wall too. It remains to be seen how the "fixed" chipset will affect this tho.

The Chipset.. or more like the SB is only connected via DMI, there is no QPI link between the cpu and the SB.

Afaik there isn't even a QPI link anymore on consumer SB, on arrandale the QPI link was connecting both dies, now there isn't such a thing anymore.

Imho the wall you mentioned is releated to the L3 cache, people forget that with SB the L3 runs @ the same clock as the core.... Westmere also reaches ~5-5,2ghz L3 clock.

So i doubt with a new chipset stepping you will see significant improvments on the chips.

Anyway thats quite offtopic.