SEARCH
ASUS Member ID
Password
Forgot password?
Join member

ASUS Support Home Location : Forum  
  Forum  
  Topic : [Problem]ECC doesn't work with P8B WS? New Topic
Jamon
Posts : 4
Level :
Tech Points : 2
From: United States
Posted:9/5/2011 2:07:00 PM # 1
With BIOS 0605, using ECC RAM in the P8B WS, booting showed the message:
"ECC function will be enabled automatically when user installs CPU and memory with ECC function"
It was not detecting and utilizing the ECC function. I found others with the same problem:
"Asus advertises this as supporting ECC memory and so does Newegg, however the motherboard while being able to use ECC memory does not utilize any of the error correction functionality of ECC memory."
-- http://www.newegg.com/Product/Product.aspx?Item=N82E16813131725&Tpk=p8bws
"I have ECC memory installed and a Intel Xeon E1270 chip that supports ECC but it doesn't seem to be using it. ... Getting no responses and no support on this issue has been very frustrating."
-- http://vip.asus.com/forum/view.aspx?SLanguage=en-us&id=20110615074025320
"I wonder if there’s been any sign of actual usage of the ECC memory functions your memory comes with in conjunction with C206 chipset & XEON cpu. So far, I have yet to see such a sign on my configuration, which is a XEON E3-1275 plus 4x4GB ECC ram @1333MHz from Crucial."
-- http://www.servethehome.com/asus-p8b-ws-sandy-bridge-xeon-lga1155-c206-motherboard-review/
I emailed Asus on August 17th, 2011 asking them to add ECC support for KVR1333D3E9SK2. On August 25th a new BIOS was released:
P8B WS Formal BIOS 0704 release.
Add "ECC Configuration" item in BIOS "Ai Tweaker" when using ECC DIMM.
"KVR1333D3E9S" was also added to the memory qualified vendor list, so I assumed it was fixed.

I was excited and installed the new BIOS. The message was gone, and there was a new menu item that had the options: "Auto", and "Disabled".

I loaded Memtest86+, and it still said no ECC functionality. I also ran 'dmidecode' in Linux like the NewEgg reviewer, and it still said no error correction.

I've tried to find ways to check ECC status, but cannot find anything. There are some drivers for Linux, but none seem to work with Xeon Sandy Bridge Cougar Point chipsets.

I went to ask Asus through the tech support form, but it required huge amounts of info, from motherboard serial number, to VGA vendor and model. It was too much. Even just to post on this forum Asus required birthdate, postal address, telephone number, etc. Why?

There must be others out there with the same problem, so I'm posting what I know, and what the problem seems to be. Maybe someone has a solution.

Currently, I cannot see any indication ECC is active and able to correct errors. If it is, I do not know how many errors were corrected, or even that it's working at all.

Thank you.

  Reply Reply Quote & Reply Quote & Reply
arnassan
VIP Member
Posts : 10
Level :
Tech Points : 7
From: France
Posted:9/10/2011 8:00:00 AM # 2
Hello.
You did some nice trouble shooting for asus...I never managed to get
an ansmer ...

Some things , I have found :

-there are real "Server/Workstation" motherboards ,
 theses use ECC(they are different with green pcb , small radiator ...)

->go to this website , you will find some  thorough test:  ServeTheHome.com
LGA 1155

The  X58 or C206 (P6T7WS or P8BWS) are not the "real thing":
 - a raid controller will not work in a pci-express connector
 - ethernet will (not from Intel) loose contact
 - Ecc ...?
 - some parts don't work with professional software , there is no
 driver for Linux .... and so on
 -onboard raid ( from realtek ) is  :- (
So , the solution might be to dump the P8BWS , and get some real c206    
  or C204 or C202 server/workstation motherboard .

  Reply Reply Quote & Reply Quote & Reply
Fred
Posts : 4
Level :
Tech Points : 2
From: United States
Posted:10/10/2011 6:30:00 AM # 3

Quote:
Author: Jamon Posted: 9/5/2011 2:07:00 PM

I was excited and installed the new BIOS. The message was gone, and there was a new menu item that had the options: "Auto", and "Disabled".

I loaded Memtest86+, and it still said no ECC functionality. I also ran 'dmidecode' in Linux like the NewEgg reviewer, and it still said no error correction.

I've tried to find ways to check ECC status, but cannot find anything. There are some drivers for Linux, but none seem to work with Xeon Sandy Bridge Cougar Point chipsets.

I went to ask Asus through the tech support form, but it required huge amounts of info, from motherboard serial number, to VGA vendor and model. It was too much. Even just to post on this forum Asus required birthdate, postal address, telephone number, etc. Why?

There must be others out there with the same problem, so I'm posting what I know, and what the problem seems to be. Maybe someone has a solution.

Currently, I cannot see any indication ECC is active and able to correct errors. If it is, I do not know how many errors were corrected, or even that it's working at all.

I followed essentially the same path.  Disabled or "Auto" is not terribly reassuring.  Memtest86++ doesn't add any confidence, nor did the response from Asus tech support.  I'm not a happy camper ...

So ... Here's the plan.  I pulled out the Intel Xeon E3-1200 data sheet (volume 2).  PCI device zero has lots of CPU configuration data.  Use "lspci -s 0:0.0 -xxxx".  From the data sheet (Table 2-7), offset 0x48-0x4F specifies MCHBAR ... in my case, the value was 0xFED10001 (YMMV).  (The trailing one indicates that mapping is enabled).

 Since "dd if=/dev/mem" would not cooperate, you write a short program using mmap on /dev/mem to access the MCHBAR area and write it (32KB) to a file.  Then examine hexdump of that file.  As per tables 2.16.2 and 2.16.3 from the data sheet, we examine offsets 0x5004 and 0x5008 in the hexdump.  In my case, both values were: "10 10 66 03".  All of the bit fields seem reasonable.  In particular, the "03" is specified as "ECC active in both I/O and ECC logic" !

Now, I have not read the entire data sheet, and there could be more to it ... but it looks to me like the Sandy Bridge Xeons perform ECC operations in the processor itself, and the above referenced bits SEEM to be turning on this capability.

My GUESS is that memtest86++ has not yet been updated to support Sandy Bridge Xeons.  Lots of time/anguish wasted on this issue ... guess that's why they call it the bleeding edge.  Hopefully I can save someone else some trouble ... enjoy ..




  Reply Reply Quote & Reply Quote & Reply
Jamon
Posts : 4
Level :
Tech Points : 2
From: United States
Posted:10/18/2011 3:14:00 AM # 4
I bought a new motherboard and RAM:
  • Super Micro X9SCM
  • Crucial DDR3 1333 (PC3 10600) ECC Unbuffered
I thought maybe it just didn't like the Kingston, but if not, I had another motherboard to use for now.

I put the Crucial in the P8B WS, and it was the same. There was no indication anywhere of ECC; not in BIOS (it doesn't even say "Enable", only "Auto"); not in memtest; not in dmidecode.

:(

There have been 5-star ratings on newegg recently, with reviews that dismiss the ECC problems.

As for ECC, the dmidecode output from a previous review is likely not relevant, since the memory controller is on the processor (and not on motherboard). Version 0704 of the BIOS has an option to enable/disable the ECC support.

ECC support fixed w/ BIOS update

ECC support was added via BIOS update 20110825:

ECC memory support ( not a big deal with today's RAM)

It's sad because with everyone pretending there's no problem, Asus has no motivation to fix it. My poor motherboard is sitting there, unused, collecting dust and cobwebs.

I bought this motherboard specifically because it had ECC. If it didn't, I could've chosen one with more features for desktop chips. I don't just want a word "ECC" on the box, I want to see the corrected errors and know when it's working.

Even if it is correcting errors, and your long process to verify that is correct, what use is it if you can't know when there were errors? I can't trust some values in a hexdump. Unless it changes to reflect corrected errors, I will never feel confident that it is doing anything at all.

So is the case that Sandy Bridge does error correction in the processor, so there's no interface or indication in the software world?

From my Super Micro, with either the Kingston or Crucial, there is indication.

# dmidecode 2.10
SMBIOS 2.7 present.

Handle 0x0027, DMI type 16, 23 bytes
Physical Memory Array
	Location: System Board Or Motherboard
	Use: System Memory
	Error Correction Type: Single-bit ECC
	Maximum Capacity: 32 GB
	Error Information Handle: No Error
	Number Of Devices: 4

Handle 0x002A, DMI type 17, 34 bytes
Memory Device
	Array Handle: 0x0027
	Error Information Handle: No Error
	Total Width: 72 bits
	Data Width: 64 bits
	Size: No Module Installed
	Form Factor: DIMM
	Set: None
	Locator: DIMM_1A
	Bank Locator: BANK0
	Type: Unknown
	Type Detail: Synchronous
	Speed: Unknown
	Manufacturer: A1_Manufacturer0
	Serial Number: A1_SerNum0
	Asset Tag: A1_AssetTagNum0
	Part Number: Array1_PartNumber0
	Rank: Unknown

Handle 0x002D, DMI type 17, 34 bytes
Memory Device
	Array Handle: 0x0027
	Error Information Handle: No Error
	Total Width: 72 bits
	Data Width: 64 bits
	Size: 4096 MB
	Form Factor: DIMM
	Set: None
	Locator: DIMM_2A
	Bank Locator: BANK1
	Type: <OUT OF SPEC>
	Type Detail: Synchronous
	Speed: 1333 MHz
	Manufacturer: Kingston        
	Serial Number: 8B028C09  
	Asset Tag: A1_AssetTagNum1
	Part Number: 9965413-028.A01LF 
	Rank: 2

Handle 0x0030, DMI type 17, 34 bytes
Memory Device
	Array Handle: 0x0027
	Error Information Handle: No Error
	Total Width: 72 bits
	Data Width: 64 bits
	Size: No Module Installed
	Form Factor: DIMM
	Set: None
	Locator: DIMM_1B
	Bank Locator: BANK2
	Type: Unknown
	Type Detail: Synchronous
	Speed: Unknown
	Manufacturer: A1_Manufacturer2
	Serial Number: A1_SerNum2
	Asset Tag: A1_AssetTagNum2
	Part Number: Array1_PartNumber2
	Rank: Unknown

Handle 0x0033, DMI type 17, 34 bytes
Memory Device
	Array Handle: 0x0027
	Error Information Handle: No Error
	Total Width: 72 bits
	Data Width: 64 bits
	Size: 4096 MB
	Form Factor: DIMM
	Set: None
	Locator: DIMM_2B
	Bank Locator: BANK3
	Type: <OUT OF SPEC>
	Type Detail: Synchronous
	Speed: 1333 MHz
	Manufacturer: Kingston        
	Serial Number: 8B027F09  
	Asset Tag: A1_AssetTagNum3
	Part Number: 9965413-028.A01LF 
	Rank: 2

It says right there, "Single-bit ECC". This is with the same RAM.

The Asus P8B WS is a tweaked desktop board pretending to be workstation class. I've seen no trustworthy standard indicator that it is correcting errors. Even if it is, I want interfaces so I can know when it's correcting them.

Until then, it is a disappointment.

  Reply Reply Quote & Reply Quote & Reply
Kenny
VIP Member
Posts : 1461
Level :
Tech Points : 1659
From: United States
Posted:10/18/2011 5:17:00 AM # 5
I think there was a discussion about how the P8B WS did not support the use of ECC memory with client OS, such as Windows 7, however you decided that was not true in your case. Now you say it won't correct errors.

  Reply Reply Quote & Reply Quote & Reply
Fred
Posts : 4
Level :
Tech Points : 2
From: United States
Posted:10/22/2011 11:56:00 PM # 6

Quote:
Author: Jamon Posted: 10/18/2011 3:14:00 AM
The Asus P8B WS is a tweaked desktop board pretending to be workstation class. I've seen no trustworthy standard indicator that it is correcting errors. Even if it is, I want interfaces so I can know when it's correcting them.

Until then, it is a disappointment.


I am in complete agreement with your conclusion.  A BIOS update is desparately needed to add accurate ECC configuration information to the DMI data provided to the operating system ... and this should have been completed a long time ago.

However, the fact (or perhaps supposition) that error correction is performed is valuable to me by itself, even if it is not reported through the operating system.  I admit that the evidence provided above is not terribly compelling.  I actually considered severing one of the bit traces to provide a more definitive analysis ...

In my application, I really need two PCI-E x16 (physical) slots ... so my alternatives are rather limited ...


  Reply Reply Quote & Reply Quote & Reply
Gardas
VIP Member
Posts : 4
Level :
Tech Points : 1
From: Czech Republic
Posted:2/14/2012 4:37:00 PM # 7
Hello,
has anything changed w.r.t. ECC support and error reporting to OS on P8B-WS board? I'd like to purchase one with proper ECC (SECDED) support and P8B-WS caught my eye. Unfortunately if the issue is not solved yet, I'll need to go to competition after having a lot of Asus boards in the past. :-(
Thanks!
Karel

  Reply Reply Quote & Reply Quote & Reply
Gardas
VIP Member
Posts : 4
Level :
Tech Points : 1
From: Czech Republic
Posted:5/16/2012 9:04:00 PM # 8
Looks like Asus is preparing P8C board with updated chipset and support for Ivy Bridge Xeons. The crucial question for me is: is ECC support implemented well that means better than on P8B? Does BIOS supply right information to the OS so OS is able to tell you which module is faulty and if ECC is in use at all?
Thanks!
Karel

  Reply Reply Quote & Reply Quote & Reply
Gardas
VIP Member
Posts : 4
Level :
Tech Points : 1
From: Czech Republic
Posted:5/24/2012 6:17:00 PM # 9
I've quickly gone thorough P8C board manual and have not found any ECC related option(s) in its BIOS description. The whole thing looks like Asus screwed ECC options while transiting to GUI-based BIOS and now as ECC looks like low-priority to them they add ECC related options just very slowly back. I'm quite sick of the situation where workstation board gets more attention for its overclocking capability than for its work-related capabilities like ECC certainly is.
Asus, please recover itself! Thanks! Karel

  Reply Reply Quote & Reply Quote & Reply
1
Previous Topic:: Next Topic
Feedback for our Website ModeratorModerator: New Topic Related FAQ Related FAQ Forum Home Forum Home
Forum Model :
Legal Information | Privacy Policy | Sitemap | Contact Us | ASUS Global
  ©ASUSTeK Computer Inc. All rights reserved.