Sandy Bridge problems

[Stereodude]

And you cannot put in a $15 PCIe controller. :skepo:

Why should I? I still plan to swap out the motherboard. Microcenter will accept them back until May 15th offering a full refund in the form of a gift card / store credit. I'm not sure if I want to do a 1:1 swap or take the refund and get a Z68 chipset board after they come out May 8th.

 

[BingBangBop]

It should be easily testable. Just switch a working drive to the SATA port of the failed Hitachi drive. If it also fails then it is likely the port. If the Hitachi works in one of the high speed SATA ports then again it is the port but if it doesn't then it is likely the drive.

I doubt that if the port failed, it will be destructive to the working drive.

 

[Stereodude]

The same drive seems to be working on the same port now. I copied a pile of data off the drive after rebooting to another (on a SATA-III) without issue. :scratch:

I guess I will stress test the drive / port some more later.

 

[LiamC]

Why do I have the uneasy feeling that Intel may have been lying about the frequency of the bug? The PR story was 5-15% over a number of years, now commonly assumed to be just 5%. But my local supplier told me that they had had heaps of returns with actual failures.

Because the story Intel told coincided with a financial analyst day, and it seems to have been designed to assuage them. If the the failure rate was as low as Intel claimed, then I'd be surprised that they caught it in their accelerated testing so soon. If the failure rate was high enough to take a billion dollar hit to earnings (and they did), then the failure rate in the press release/official statement looked suspiciously low.

When I first read the official line, it had week-old fish all over it.

//yours in paranoia

 

[Stereodude]

Well, it happened again. This time I decided to try the optical drives connected to the other 3.0Gbs SATA-II ports also after the HD generated a delayed write error. Sure enough they were non responsive / unreachable. So, I feel safe saying that this is the Sandy Bridge chipset error.

 

[time]

I thought Intel said it would just slow things down?

This isn't helping my paranoia.

 

[Stereodude]

Well, it's basically dead. Even after powering off the system and rebooting or warm rebooting pretty much any attempt to use one of the devices connected on the 4 SATA-II (3.0Gbs) ports will cause the HD Light to stay on and the drive access to fail / error out. Even running chkdsk on a almost empty drive is too much activity. I pulled the Hitachi drive and was able to access the data without issue in another PC though.

I think Intel's estimates are a tad low / optimistic. :rotfl:

 

[Handruin]

Maybe time for a Z68 board?

 

[Stereodude]

Maybe time for a Z68 board?

Maybe... Not sure if I want to swap this out now, or wait until May 8th and see what Z68 boards pop up on the market. Microcenter says I have until the 15th to return the board for replacement / refund. I'm tempted to take the refund and get a Z68 board though I'm not sure I want to be an early adopter on the Z68 like I was on the P67.

 

[Handruin]

I understand the hesitation. Given the incredible problems, bad press, and monetary loss of the P67, I'm hedging a bet that the Z68 went through a few more QA cycles before its release. :-D They can't stand to take another huge hit this year. This may be the best board to buy after a huge problem they just went through. Then again, I'm only making a guess here.

 

[Adcadet]

My understanding is tha this $1 billion mistake was the result of one engineer making a late change to the board after most testing was complete. I wonder if she/he still has a job.

 

[Stereodude]

I find it hard to believe that a single person managed to make a late change to the chipset all by themselves. Doesn't Intel do peer reviews? :scratch:

At this point I'm leaning toward at least waiting to see what the Z68 boards look like and how much they are before getting a replacement for this board.

 

[Mercutio]

Doesn't Intel do peer reviews?

First thing that went through my mind, too. Especially since so much modern development work depends on team-based practices.

 

[Stereodude]

So I decided to put together a little timeline on the P67 experience. I bought it 1/27/2011 and as of 4/27/2011 all 4 SATA-II (3Gbs) ports are basically dead.

The system was not heavily used. It was not a primary use machine. It was used basically for video compression. Up until 4/15 the SATA-II (3Gbs) ports were only used for optical drives (burning some discs). There was a SSD and SATA HD on the SATA-III (6Gbs ports) from the beginning. On 4/15 I received 5 Hitachi Deskstar 7K3000 HDS723020BLA642 2TB 7200 RPM HD and decided to full format each of them (one at a time) using one of the SATA-II ports. After finding out that 2 of the 5 were bad (one has 224kB of bad sectors another does the whir click / whir click / whir click routine after attempting to format it) I decided to leave one of the good 7K3000's connected to use it a little (doing x264 video compression) to see if the HD would fail or not. And, voila! a whopping twelve days after first connecting a HD to the controller the SATA-II ports are effectively dead.

So, apparently full formatting 5 2TB HDs (4 really since one failed at 0%), doing a full surface scan chkdsk of one 2TB HD, burning ~100 BD-R discs, and a little video compression on an attached HD is enough usage to have the Cougar Point bug rear it's head.

 

[Mercutio]

I notice that the warranty terms on almost all the new P67 motherboards, even $200+ products, are one year. I don't know if that's the result of a long term policy change from the usual gang of idiots or just an early adopter penalty but it makes an upgrade substantially less appealing.

 

[LunarMist]

It's probably due to being cheap crap. Of course they don't reimburse you for hundreds or thousands of dollars in replacement software.

 

[Stereodude]

I notice that the warranty terms on almost all the new P67 motherboards, even $200+ products, are one year. I don't know if that's the result of a long term policy change from the usual gang of idiots or just an early adopter penalty but it makes an upgrade substantially less appealing.

Gigabyte is still offering a 3 year warranty on their motherboards.

 

[Mercutio]

I looked at a couple Gigabyte, Asus, MSI and Asrock boards on Newegg and it was reporting 1 year warranties last night until I hit $225 boards.

 

[Bozo]

It's probably due to being cheap crap. Of course they don't reimburse you for hundreds or thousands of dollars in replacement software.

Why would you have to by new software if your motherboard dies?

 

[ddrueding]

Why would you have to by new software if your motherboard dies?

Officially, MS OEM software is tied to a motherboard. If you lose the motherboard, you lose your software license.

 

[Stereodude]

I looked at a couple Gigabyte, Asus, MSI and Asrock boards on Newegg and it was reporting 1 year warranties last night until I hit $225 boards.

Well, I didn't look at Newegg, I looked at the manufacturer's website.

 

[Mercutio]

I wound up buying an Asus board yesterday because Gigabyte isn't answering my RMA requests and Intel's boards don't have enough onboard eSATA ports for me.

I'm going to build a machine that can actually use all 20 drive bays in one of my big rack cases.

 

[Stereodude]

So, I have my Gigabyte GA-P67A-UD3P in the car today and plan to drop by Microcenter this afternoon after work and swap it out for a Gigabyte GA-P67A-UD3P-B3. I only paid $139.99 for it and I see the GA-P67A-UD3P-B3 is now ~$170 at the cheapest (even online), and the Z68 boards are rumored to be at a 10-20% premium over the P67 boards.

Even if they're the same price I can't see spending another $30 for the two distinguishing features that are basically useless to me (the SSD caching & Quick Sync). :scratch:

 

[Mercutio]

I bought an Asus P8P67 Deluxe, mostly because it has enough SATA ports. If I hate it in the way that I usually hate Asus boards, I'll sell it to some other sucker.

 

[Stereodude]

I added a Dell Perc 5i that I had sitting in my spare parts bin in the basement to my P67A-UD3P to run two 2TB RAID-1 arrays (assuming I can get 4 working drives from Hitachi) since the P67A-UD3P only has 6 SATA ports whereas my P35 and P45 based Gigabyte boards all had 8 onboard SATA ports.

 

[Mercutio]

I'm updating my gaming rig. I don't particularly want to stick an expensive disk controller in that machine but it has three 3-in-2 3.5" drive bay adapters in it that I can use for occasional bulk storage in case I'm doing something silly with my file servers.

The math going on in my head is that I'm going to buy an expensive motherboard anyway, so I'd rather pay ~$30 more to get the extra SATA ports than $100-ish for another PCIe card.

 

[Adcadet]

Does the Pro not have the same number of SATA ports? I like my P8P67 Pro.

 

[Mercutio]

The deluxe one has more SATA 6Gpbs ports as I recall. And some spiffier heat sinks.

 

[BingBangBop]

Isn't the z68 chipset about to be released (May 11th?). Why would anyone want the current crop of Sandy Bridge MB's with their old outdated chipsets in preference.

 

[Stereodude]

Isn't the z68 chipset about to be released (May 11th?). Why would anyone want the current crop of Sandy Bridge MB's with their old outdated chipsets in preference.

You have some information about how much better the Z68 is when compared to the P67 that no one else is aware of? :scratch:

 

[Mercutio]

Honestly, if the differences are that big, I'll probably just buy another one and find a home for the older one. I doubt it will be that big or that the discounts for the older stuff will be meaningful. Older i7 motherboards are still plenty damned expensive.

 

[Mercutio]

I installed my new 2600k today. A few quick thoughts:

1. Asus has a GUI BIOS that I found more than a little annoying. Award and Phoenix have been the same for years and there's no reason to show me moronic little gauges.
2. The 2600k has a tiny, weak fan and average temperatures on my 2600k are higher than I'd like, idling at around 44C without me screwing around with anything. It's also insultingly loud, meaning that I can hear it. This must be fixed.
3. Really, Asus? None of the included drivers will run their Setup.exe under Server 2008R2, so I got to load them all through Device Manager. The Intel Ethernet Adapter and Serial port (yes, serial port) still won't install.
4. Plus sides: Clear CMOS on the rear panel and hard buttons for Power On and Clear CMOS on the inside, Comprehensive manuals, little 3.5" drive with a couple USB3 ports on it.

 

[Bozo]

Honestly, if the differences are that big, I'll probably just buy another one and find a home for the older one. I doubt it will be that big or that the discounts for the older stuff will be meaningful. Older i7 motherboards are still plenty damned expensive.

Because of the mess in Japan, Ive seen prices on some items starting to rise. Low end MSI video cards were ~$40.00- now they are closer to $60.00 and the available models has deminished.

 

[Mercutio]

I need to figure out what I want to do about cooling my 2600k. I'm given to understand that they should be cooler than previous i7s, but mine seems to have stabilized at about 41C at idle, and I'm not doing any overclocking on it yet.

 

[ddrueding]

I'm getting good performance with a Zalman tower. Only drawback is that you can't swap the fan. Not that the one it has is bad or noisy, I just prefer having control.

i7-970 @ 3.9Ghz including some overvolt, 100% load @ <50C.

 

[BingBangBop]

Whenever, I get consistently bad temps (w/o OC'ing) with a factory fan+heatsink then it has always been the thermal grease/paste application (i.e. user error). However other alternatives are that your heatsink isn't flat or that you didn't tighten it down properly.

 

[Mercutio]

I know my heat sink is on properly and I didn't molest the heat pad prior to installation, but that's a distinct issue from the noise level anyway.

I'm thinking I might go with a Scythe Kabuto. I think I have to dust the inside of my case less often when I have a CPU fan that mounts horizontally.

 

[ddrueding]

Looks like a good unit. I prefer horizontal airflow because the motherboards are mounted in my case such that GPU and CPU exhaust air is all directed upwards.

 

[CougTek]

If money is no object and you have plenty of space, the Noctua NH-C14 is more efficient and can therefore allow a lower fan rotation speed for similar or better cooling results. It's shorter too (105mm) if you only use the bottom fan. It's a 70$ heatsink though.

 

[ddrueding]

Wow that is a big unit with both fans going...