Intel Identifies Sandy Bridge Chipset Design Error, All Shipments Stopped

[15th Warlock]

Looks like it's affecting only the 3Gbs SATA ports, good thing my mobo has 6 other 6Gbs SATA ports (2 from P67 and 4 from dual Marvel controllers) which are backwards compatible with my SATA 2 drives, so I'll just move my drives to those ports in the meantime.

Gigabyte (and all other mobo manufacturers as of this moment) has not addressed this problem yet, so I'm gonna hold my horses until I receive any info from them, no point in going the RMA route as all mobos in the channel must have the same issue.

Will wait for Intel to update their chipsets and distribute them, then I hope GB will do he right thing and recall their p67 mobos.

IMHO Intel did the right thing by informing their user base about this problem, and doing a recall, I applaud them for that.

All hardware companies have had bad products batches in the past, it's impossible not to, as complexity increases for these parts. I'm glad Intel is addressing the issue and will take care of its costumers, to me that's all that matters in the end

 

[bear jesus]

If Intel was able to reproduce the issue with stressing the chip then surly high usage would bring the issue out sooner?

If say you run the computer 24/7 with near constant hard drive usage would that not speed things up?

I think the main problem is killing resale value and looking at the sale forum most of us like to sell on our old hardware.

 

[Solaris17]

I'll take a cheap faulty in 2 years 1155 board..lol just use sata 6 when your 3.0 SATA controller dies FTW..




Just 3.0 SATA controller not the 6.0 from what i've read.

when in doubt sata card it they all have a ludicrous amount of pci-e slots now anyway.

 

[Mindweaver]

when in doubt sata card it they all have a ludicrous amount of pci-e slots now anyway.

I see a 3ware 9650SE in my future..

 

[Hayder_Master]

that's cuz they want seal people with new socket which is no different from 1156 chipsets, shit on u INTEL

 

[arroyo]

It looks like that:

Intel = great CPUs + bad chipsets (P55,P67,H67)
AMD = medicore CPUs + great chipsets (890GX, 890FX)

I would like to see a misalliance like this:
Intel 2600K + AMD 890FX board = THAT WOULD BE AWESOME

 

[cdawall]

It isn't so bad, Intel is accepting RMAs on the affected processors. This is why I never buy the first revision of a processor release. Though to be fair, this isn't an issue with the processor, it is a problem with the integrated chipset.



It isn't like AMD hasn't had some pretty big issues with their silicon...TLB anyone?

at least the TLB bug didn't require me or you to take the said affected part go offline for x number of days to ship it all the way back to the manuf. wait X number of days for them to go oh this was part of that bad batch wait X number of days for it to show back up on your doorstep to get reinstalled into your PC. so instead thanks to intel any of these that have already found there way into a server just went offline because intel fucked up. this is not a little thing this is a big thing letting bugged parts out is an issue that is way there is a pending lawsuit against Nvidias flawed chipsets/IGP's.

the TLB bug required a patch readily available on AMD's website and later built into windows. not that big of a deal for a server environment chipset taking a shit on you BIG DEAL.

 

[meran]

wow i was supposed to buy this shit

 

[LAN_deRf_HA]

Why did newegg just take down the SB procs? I get the boards, but not the cpus.

 

[Mindweaver]

Why did newegg just take down the SB procs? I get the boards, but not the cpus.

Yea, but what good is a processor with out a MB? I would say they did this to help there Customer support team.

They are throwing a party at the AMD camp! I'm not laughing to hard i have a P67 board that arrived today..

 

[LAN_deRf_HA]

When amd had a problem with their actual processors I don't recall them being taken down. Shit makes no sense. And newegg shouldn't care that they can't sell boards with the cpus. If someone wants to buy a cpu let them, this is just a needless loss in sales.

 

[Mindweaver]

I never said that it was right. Maybe they are looking back at that very problem and want to handle it different with this problem. idk?

 

[niko084]

Read about this earlier today...

Kinda sad I just ordered a 2600K and an ASUS P8P67 PRO...
Guess I'll have to see and if it starts to screw up I'll warranty it :-/

 

[meran]

Read about this earlier today...

Kinda sad I just ordered a 2600K and an ASUS P8P67 PRO...
Guess I'll have to see and if it starts to screw up I'll warranty it :-/

after destroying all your hdds

 

[KieX]

From UK retailer's message about the news:

Motherboards with devices connected to Ports 0 & 1 remain unaffected.

source

So on most motherboards, that is 2 intel ports plus 2 marvel ports (unaffected)... if I count properly you need 5 or more hard drives to have a small chance of developing a fault over 3 years (estimated).

I don't see why the fuss.

 

[PVTCaboose1337]

This reminds me of the Intel Pentium chip problem way back. It is almost undetectable to the end user. Only a few can notice it. Glad to see Intel is doing the right thing, unlike last time this happened. Last time they denied it.

 

[niko084]

after destroying all your hdds

Doesn't say anything of the sort simply degrading performance.
Don't make it sound to be some epic huge unrecoverable problem.

 

[PaulieG]

after destroying all your hdds

Read the information in a little detail, it will not destroy drives. It will degrade performance. So, no real risk to hardware. It's also unclear as to what percentage of boards may experience this problem at all. I'm gonna hold on to my chip and board, and not overreact because I'm happy with it. When March or April comes, and the board partners decide how they are going to handle recalls, then I will probably send it in for replacement.

 

[Psychoholic]

I am glad i purchased mine when i did, seeing as how newegg took them down.

Should anything happen, it will just give me an excuse to move to SATA 6GB

 

[LAN_deRf_HA]

I'm bugged mostly by resale value decrease. I'm fine with using the sata 6 if I even need to. Be nice if they could find some way to test if your particular board will be one of the 15% possibly effected after 3 years. Man this really does sound like a minor issue.

 

[TheGuruStud]

It isn't like AMD hasn't had some pretty big issues with their silicon...TLB anyone?

You mean the bug where you had to be at 100% load and all cores virtualized with pigs flying overhead for the error to occur?

Yeah, whew, catastrophe averted getting that TLB fixed. That was all intel hype.

 

[LAN_deRf_HA]

Explanation of the exact cause of the issue http://www.anandtech.com/show/4143/the-source-of-intels-cougar-point-sata-bug

Wonder if a bios change could fix it, though they'd have probably suggested it already. Undervolting may be sufficient to indefinitely hold it off.

 

[cdawall]

Explanation of the exact cause of the issue http://www.anandtech.com/show/4143/the-source-of-intels-cougar-point-sata-bug

Wonder if a bios change could fix it, though they'd have probably suggested it already. Undervolting may be sufficient to indefinitely hold it off.

after reading that undervolting wouldn't do anything the bias is fucked up and now its leaking they need to replace the gate.

If you have a desktop system with six SATA ports driven off of P67/H67 chipset, there’s a chance (at least 5%) that during normal use some of the 3Gbps ports will stop working over the course of 3 years. The longer you use the ports, the higher that percentage will be. If you fall into this category, chances are your motherboard manufacturer will set up some sort of an exchange where you get a fixed board. The motherboard manufacturer could simply desolder your 6-series chipset and replace it with a newer stepping if it wanted to be frugal.

that sounds pretty shitty use more than port 0/1 and you could end up with just ports 0/1...as for the people asking earlier guess that means all of the overclockers pushing these boards will never notice

 

[Thefumigator]

Just some observations...
What about making them work as SATA 1.0 (150m/s), it may be slow for some SSD, but still enough for most regular HDD and DVD rewriters of course... Not sure if its possible, I've never seen on bios cmos setup an option to turn your SATA 2.0 connectors to SATA 1.0 (maybe on windows device manager...). Also not even sure if this could help the chipset.

Buggy chipsets... what's next? buggy VGAs? Not that I found ok to see buggy hardware around, but I believe that is not big deal, if for some reason I get something buggy I would just solve it with a workaround... my Phenom 9500 worked ok for me, and that intel mobo sounds like an amazing deal if e-tailers do price drop for clearance. I would fix it with a pair of cheap SATA controllers I have sitting somewhere...

 

[bear jesus]

One thing that is confusing me is everyone quoting 3 years as a definitive thing, if users have been having issues reporting it to Intel and then Intel reproducing the problem all within a month that would suggest to me if you have drives under heavy use on those ports it will wear them down pretty fast.

I know I'm far from a good example and not even using the chip in question but my PC runs 24/7 and is always either reading or writing to the primary drive which i always put on the first port of the on chip controller, would that mean if i had given in and bought something from the sandy bridge line i would be at a higher risk of wearing out the connection than most normal users?