Would you still risk doing a Sandy Bridge build?

[PingSpike]

Probably not. At around $100 for a basic H67 board that is pretty expensive price to pay for a busted board that will be worth not much on the used market due to its condition. And it appears a few places that are still selling the boards seem to be raising prices :| probably due the shortage from more reputable places yanking their stock. Sure you can buy a controller card...but those are like $30 for two ports right?

Assuming the board maker offers a replacement I'm probably looking at computer downtime while the RMA ships...and lets be honest unless they're really great and do advanced shipments (probably not) their RMA department is going to be overloaded and I'll be stuck holding my dick for maybe weeks. Its unlikely your board would fail for another reason, but if it does you're fucked because there isn't any replacement stock, bleh.

If it was the summer wouldn't be so worried about computer down time I might not care, but not right now. If I'd already bought the system and built it I'd just plug it into the sata6 port and wait and see. If I hadn't opened the boxes yet or gutted my old PC though I'd send them back myself. And I hate sending things back.

And I know some people are really taking that 15% number from intel at face value...but I agree with Spook. I have no reason to believe that isn't an optimistic statistic that is actually probably just and educated guess. And you know what is worst about that? 15% over 3 years is still pretty damn bad even if you believe its 100% accurate.

 

[Anomaly1964]

As long as I can play Crysis 2 when it comes out, I'll be fine...GRIN

 

[Edrick]

I only have 2 drives in my system and they are on the SATA 6Gps ports. So this bug will not affect me at all. So no panic here, and yes I would build another SB system if I did not use more than 2 drives.

As for resale value, well, when I upgrade to LGA2011 in Q4, I will simply RMA my current MB at that time, then sell it as a "New" board and make back most of my money.

Sure this news was a little disturbing at first, but now that I know the facts, it is only a small bump in the road. And I am still loving my SB machine now.

 

[Avogadro]

I built a SB system for myself two weeks ago so the "go/no-go" decision has already been made. Sure, with the benefit of hindsight I should have waited. However, I knew that I was buying brand new technology, and problems like this are a risk with early adoption. Granted, this is way more serious than BIOS updates or waiting for some new drivers, but still, it's not like the situation itself should be a surprise. Magnitude of it: for sure.

All I can do now is make the best of the bad situation. Fortunately, I have the benefit of these forums, so I can hear what the experts are thinking and doing. That's really helped me not to worry. When I get home tonight, I'll open up my computer, switch my HDD and DVD burner to SATA3 ports (the SSD is already on one), and make double-extra sure my mobo registration info is complete.

At some point Intel and Asus will figure out what to do about all this, and Asus will send me an e-mail saying "hey d00d yer mobo is borked" and tell me how to get a replacement. Dealing with some computer downtime will suck (those daily quests won't do themselves after all ), but maybe I'll be able to do a replacement exchange at a local retailer instead of by mail.

Now, I can afford to be patient because my computer isn't mission-critical. If someone built a SB machine that is, he/she has every right to be angry about this.

But, no, I don't feel like a sucker, or that Intel screwed me (it's not like they WANT to sell you broken chipsets after all). So I'm OK with my decision.

 

[Hogan773]

I have a little concern about sending my mobo, which is working just fine, back to some place where a dude with a soldering iron is going to rip out one chip and plop in another. Maybe it will make more sense to just send everyone a brand new board that was built on the automated assembly line. I'd rather not have my board have surgery from some hack with a pile of 3000 mobos sitting there staring him the face.....you think he's gonna be careful?

 

[Diogenes2]

This. Not a snowball's chance in hell I'd build a Sandy Bridge system right now. My data is too precious to me to even risk a 1% chance of corruption, much less 5-15%.

Your precious data would be no more at risk with SB than it is right now..

If it's not backed up, it's even more at risk than my data is...

P.S.
For those still looking to buy, Amazon shows boards and CPU's in stock ..

 

[eddietandy]

Your precious data would be no more at risk with SB than it is right now..

If it's not backed up, it's even more at risk than my data is...
..

I'm more worried about data corruption in this instance, and having corrupted data backed up does no one any good.

 

[Diogenes2]

Data corruption is not part of the problem.

How can you corrupt data that is not being transmitted?

 

[abaez]

I don't think that the port just "stops" working one day. As mentioned, the connection will degrade, and bad bits or whatever will have to be resent over and over until failure.

For me, I wouldn't care as long as the data doesn't get corrupt. If the port just fails I'm fine with just switching to another but not if my data will slowly get corrupted over time until failure.

 

[Minjin]

I don't think that the port just "stops" working one day. As mentioned, the connection will degrade, and bad bits or whatever will have to be resent over and over until failure.

For me, I wouldn't care as long as the data doesn't get corrupt. If the port just fails I'm fine with just switching to another but not if my data will slowly get corrupted over time until failure.

That's what error correction in the packets is for. It just means that more and more packets will have issues and need to be resent. The connection will gradually get slower until it finally fails. I'm guessing that most people will recognize an issue long before the connection goes away completely and I doubt there will be any corruption.

 

[nOOky]

If I had not have built mine a few weeks ago I would probably wait or go with a i7-950 etc.
Since I did do it and I have only 2 sata hdd's and an external drive which isn't affected at all I will continue to use my system and not worry at all.

 

[SparkyJJO]

I don't buy broken parts to build my PC with. The chipset is broken, so no I wouldn't do it. I'd wait until they got it straightened out, then build.

I don't care if the other ports work. For how expensive the boards are I expect them to function 100%.

 

[Engineer]

Data corruption is not part of the problem.

How can you corrupt data that is not being transmitted?

So if you were writing data and suddenly, the transistor gave out, there wouldn't be corruption? Especially if it were writing to the drive file tables?

 

[PingSpike]

LOL...I agree SparkyJJO. If this were some piece of shit $45 junk board it'd be annoying but you could always say you get what you pay for. But when the "cheap" boards with no overclocking start at $90+ and will likely need to be thrown away/sold the next time Intel comes out with a new processor you kind of expect them to actually be perfect because you certainly aren't getting things cheap!

But, such is the life living on Intel's bleeding edge.

 

[PingSpike]

The Intel engineer responsible for the mistake has been identified:
http://pics.bbzzdd.com/users/pingspike/intel.jpg

Looks like this goes to the very top.

 

[alaricljs]

No worse than any power drop the system may experience. If your file system isn't up to the task of coping with those issues I don't know what you're running because I've dumped the power on all sorts of Unix and Windows boxes with an extremely tiny number of issues turning them back on. Yes, this is while they're actively using the drive.

 

[alaricljs]

The Intel engineer responsible for the mistake has been identified:
http://pics.bbzzdd.com/users/pingspike/intel.jpg

Looks like this goes to the very top.

That's not even a man purse, that's a full blown purse working on it's way toward being matched friggin luggage.

 

[PingSpike]

That's not even a man purse, that's a full blown purse working on it's way toward being matched friggin luggage.

I believe the adjectives "man" and "glittery" are mutually exclusive when used to describe a purse.

 

[kevinsbane]

The possibility of data corruption is theoretically possible. If the controller fails exactly in the middle of a write sequence. The issue is how risky this scenario is.

5-15% chance of failure over 3 years (if you don't take advantage of the RMA when it comes in a month or two)
Degradation (decreased speeds) at first, so you know it's coming; ie, low chance of an unknown, middle of the night catastrophic failure.
Needs to fail exactly at the right time in order to corrupt data.

Anyone one single of those risks I wouldn't accept on its own. But taken together as a whole, I would be ok with it as a temporary stopgap measure.

Statistically, if I left my critical data on a SATA II port, I'd be worried too. I definitely wouldn't keep critical data there. But hey, move it over to SATA III, or buy a PCI card, or takes steps to ensure the data isn't as critical (backups...) and you're golden.

Example: hard drives are relatively unreliable. We still use them anyways. However, for critical data, we use a raid-1 array to increase reliability. Fundamentally, HDD's are still not reliable enough in and of themselves for mission critical data, but there are enough risk mitigation techniques to mean that the risk of hardware failure is a non-issue.

So if the available workarounds and risk-mitigation techniques aren't enough to reduce the risk below what you're willing to tolerate, then by all means, get rid of your SB platform and go with something more tried and tested. Just know that you're losing out on the potential performance benefits of keeping your system.

 

[ViviTheMage]

I am OK with my SB build. If it's an issue later on, i'll deal with it/RMA/new board.

 

[Diogenes2]

So if you were writing data and suddenly, the transistor gave out, there wouldn't be corruption? Especially if it were writing to the drive file tables?

The possibility of data corruption is theoretically possible. If the controller fails exactly in the middle of a write sequence. ....

No, because the HDD controller ( on the drive ) would detect it and not write it to the drive..

...Unless it failed also, but then, that would not be a fault of the P67 ...

I think people are ignoring that there are two controllers at work here; the chip set and the one on the drive.. The drive doesn't just write everything the SATA controller sends it - no questions asked.

 

[dpodblood]

If I were building a system with only 2 HDD's or could find a mother board with a Marvell controller than yes I would have no problem with a SB build. That said I wasn't planning on building for a couple of months anyhow, so I might as well wait on the boards with Z68 chipsets.

 

[Hogan773]

No, because the HDD controller ( on the drive ) would detect it and not write it to the drive..

...Unless it failed also, but then, that would not be a fault of the P67 ...

I think people are ignoring that there are two controllers at work here; the chip set and the one on the drive.. The drive doesn't just write everything the SATA controller sends it - no questions asked.

Yes that is my understanding of the purpose of error checking. The chipset controller is talking over the phone with its counterpart sitting on the HDD. When there gets to be too much static, the HDD says "Huh? I can't hear you" so they need to backtrack and repeat data until they get it right. Its not like the HDD just says "well I didn't understand that but I'll just write down what I THOUGHT I heard him say and move on......"

 

[Diogenes2]

I'll just write down what I THOUGHT I heard him say and move on.....

 

[kevinsbane]

No, because the HDD controller ( on the drive ) would detect it and not write it to the drive..

...Unless it failed also, but then, that would not be a fault of the P67 ...

I think people are ignoring that there are two controllers at work here; the chip set and the one on the drive.. The drive doesn't just write everything the SATA controller sends it - no questions asked.

I stand corrected. Is it a theoretical possibility though?