Spot Signs of SATA Performance Degradation on Cougar Point P67/H67 Chipsets

[btarunr]

Apparently there's a "slim" chance of users actually being affected by the recently-discovered chipset design error plaguing Intel's Cougar Point 6-series chipsets that drive the SandyBridge processors. Intel pegs this at 5% performance degradation caused due to data transfer errors over a period of 3 years. Users of Romanian tech community Lab 501 found out a way of detecting if your hard drives connected to the damage-prone SATA ports are already showing signs of errors.

To do so, you need the HD Tune Pro software, you can try out the Pro version. Select a hard drive connected to the damage-prone SATA 3 Gb/s port, click on the "Health" tab. HD Tune checks out various self-diagnostic parameters of the drive, including SMART values. A drive connected to an affected port should fail a number of tests as shown in the screenshot below. As a stopgap solution, you can connect the drive to one of the SATA 6 Gb/s ports available. Most motherboards should have 2 internal 6 Gb/s ports, some have additional 6 Gb/s ports driven by 3rd party controllers. It's advisable to use ports of those controllers. This stopgap fix should pull you through till April-May, when motherboard vendors are expected to have motherboards with B3 revision of the Cougar Point chipset. You can then claim replacement of your board under warranty.



View at TechPowerUp Main Site

 

[btarunr]

Ironically, rectified batches of Barcelona/Agena AMD Phenom/Opteron processors also carried the "B3" revision.

 

[shb-]

So.. Where is the screen where exact drive is connected to sata3 port and showing different smart values? Looks like bad drive to me.

 

[Swamp Monster]

When I had a bad sata cable, There was only errors on (C6) Ultra DMA CRC Error Count.
In screenshot I see faulty drive and errors caused by chipset. It looks to me that it's affecting HDD's contrary to what Intel said, or HDD used for test is bad.

 

[Zubasa]

Ironically, rectified batches of Barcelona/Agena AMD Phenom/Opteron processors also carried the "B3" revision.

So are there any bios fixes for it this round?

 

[faugusztin]

@Swamp Monster: no, the drive is not having problem. Look at those values. Spin up time of 677248117 seconds ?

This screen simply shows unreliable data transfer, which results in scrambled and stupid SMART data => sign of the errors in data transfers.

 

[Swamp Monster]

^If it damages smart data, then ok, I understand. But it reveals another problem - One can't see when HDD is actually making errors ,becaue its already showing false errors.

 

[mascotzel]

from one of the guys at lab501

It's not about the actual SMART data, but about the fact that HDTune reads it corrupted. What we know is that the problem is an excessively polarised FET transistor in the clocking circuit of the S-ATA II interface (reliable info, not sure if it's official), that only some motherboards show the SMART corruption in HDTune and only on the S-ATA II ports, and this SMART corrupted readout is sometimes associated with other type of software errors. Putting it together, results SMART corrupted readout in HDTune is a sign of early S-ATA II degradation; we haven't yet seen completely not working S-ATA II interface (as Intel says that's the final evolution of this bug), but in time they should show up.

and although they omitted the results on SATA3, they tested it and there are user screenshots in the forum topic opened there (but u need to be registered to see them )

time and more user feedback will tell

PS. I posted pictures from the topic there, but too see them u need to browse the forum, and not the direct topic in the site

 

[burebista]

My two cents here.
I have a user on another forum. This is how it looks in HDTune



And how it looks in CDI or another SMART reading software.



So I'd take HDTune results with a grain of salt.

 

[faugusztin]

Do you realize that that other software could do something like checksum test, or try to run same test few times, while HDTune takes the first result value ?

 

[Konceptz]

yeah....i'll think I'll wait before pulling the trigger on I7 2600K

 

[Andy77]

yeah....i'll think I'll wait before pulling the trigger on I7 2600K

You mean P67/H67... AFAIK 2600K is a fine CPU!

Also I wouldn't trust them so much. I've read them in the past and they did some n00b-ish mistakes and this might be just one of those mistakes.

 

[Deleted member 74752]

No idea what this means but I have lost 8 drives in total on this Asus WS Revolution mb so far on the 3 Gb/s ports. This latest install is on the 6 Gb/s port.

These tests on the 6 Gb/s port -

Followed by these on the 3 Gb/s port -

 

[buggalugs]

No idea what this means but I have lost 8 drives in total on this Asus WS Revolution mb so far on the 3 Gb/s ports. This latest install is on the 6 Gb/s port.

These tests on the 6 Gb/s port -

rickss69, You're not meant to be running that benchmark with your SSD.

Suggest you take a look at the OCZ forum:

http://www.ocztechnologyforum.com/f...rtex2-Agility2-EX-Drives-REVO-and-HSDL-drives

 

[theJesus]

Why do they recommend HD-Tune Pro to check SMART values? It's available in the free version as well; I use it all the time at work.

 

[HTC]

Has anyone tried to test with HDD Sentinel? I have this prog and it's great to detect problems with non-RAIDed (and some RAIDed ones) HDDs. Unfortunately, i'm still on socket 775 so i can't test this myself

 

[bear jesus]

I have been kind of hoping that a tech site would stress the chip to try and show the effects, if Intel received complaints then managed to reproduce the problem all in the short time since release i assumed it would pretty easy to reproduce the issue in a short time and not the years that Intel has suggested.

 

[Silverel]

http://www.bit-tech.net/news/hardware/2011/02/04/gigabyte-releases-sandy-bridge-fault-checke/

If you actually want to check the Chipset, Gigabyte has a vendor neutral solution for it.


edit- Oh man... completely disregard this as a solution. Instead consider it free LOL's at Gigabyte for their... solution.




It checks what ports are being used, and flags thet SATA-2 ones... for being in use. lol

 

[Jstn7477]

Why do they recommend HD-Tune Pro to check SMART values? It's available in the free version as well; I use it all the time at work.

You're missing out on a more updated program with more SMART values then...

 

[Mussels]

http://www.bit-tech.net/news/hardware/2011/02/04/gigabyte-releases-sandy-bridge-fault-checke/

If you actually want to check the Chipset, Gigabyte has a vendor neutral solution for it.


edit- Oh man... completely disregard this as a solution. Instead consider it free LOL's at Gigabyte for their... solution.




It checks what ports are being used, and flags thet SATA-2 ones... for being in use. lol

lol what an awesome program.

what'll be funny is if you use it and it just claims that all sata II ports are faulty, even on external sata controllers.

 

[cdawall]

Ironically, rectified batches of Barcelona/Agena AMD Phenom/Opteron processors also carried the "B3" revision.

i also found it to be a bit funny that the "cursed" B2 stepping plagued not only AMD but also intel its only funnier that they have repaired them and called them B3's

 

[micropage7]

So are there any bios fixes for it this round?

yeah, nutting. defective chip should go to the trash

 

[Imsochobo]

I have been kind of hoping that a tech site would stress the chip to try and show the effects, if Intel received complaints then managed to reproduce the problem all in the short time since release i assumed it would pretty easy to reproduce the issue in a short time and not the years that Intel has suggested.

Give a system to me in my nas.
It loads 35mb/sec 24/7.