Personal computing discussed

Moderators: renee, Flying Fox, Thresher

 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 11:50 am

In my small home office, I have a dedicated mahcine for databases. It's ~4 years old, but works well for my needs. It has 2 velociraptors for the temp and log files and two other Caviar Blacks.

Soon after building it, I wished that I used ECC instead. I'm considering upgrading with two of these unbuffered Kingston kits. Is this a good idea? Will I gain enough added protection to make this worth it?

$100 seems very reasonable for this. (By the way, this itch is partly the fault of JBI's recent thread.)

It has a 775 server motherboard and a Q9550. If that processor is a problem with the ECC, I have a very similar Xeon X3360 in another machine that I can swap.
 
derFunkenstein
Gerbil God
Posts: 25427
Joined: Fri Feb 21, 2003 9:13 pm
Location: Comin' to you directly from the Mothership

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 12:18 pm

The memory controller is what determines ECC support, rather than the CPU in this case (since the memory controller isn't on the CPU), so you should be OK to do that without a CPU swap.

As to what you gain, I dunno. It seems to me that ECC isn't worth the cost simply because if you're getting data corruption errors due to memory, you just have bad memory in the first place and replacing it with cheaper unbuffered non-ECC should fix it just as well. Someone may correct me here, though. :p
I do not understand what I do. For what I want to do I do not do, but what I hate I do.
Twittering away the day at @TVsBen
 
Evan
Gerbil XP
Posts: 411
Joined: Mon Apr 21, 2003 8:42 am
Location: San Diego, California
Contact:

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 12:34 pm

The second C in ECC stands for Correcting, so ECC memory will fix single-bit errors that non-ECC memory won't, and at least find multi-bit errors. I use ECC in my desktops and while it's probably true that high-quality non-ECC should be alright, if you're using the system for something serious like a database, I'd say go for the ECC.
 
derFunkenstein
Gerbil God
Posts: 25427
Joined: Fri Feb 21, 2003 9:13 pm
Location: Comin' to you directly from the Mothership

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 12:43 pm

The second C stands for Code, I thought. Error Correcting Code. :p
I do not understand what I do. For what I want to do I do not do, but what I hate I do.
Twittering away the day at @TVsBen
 
EV42TMAN
Gerbil
Posts: 40
Joined: Fri Jun 10, 2011 11:50 am

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 12:57 pm

small office and database.... go with the ECC memory. anyone that says you can get away with high end non ECC memory doesn't directly work with databases
MCP MCDST MCSA MCTS MCITP
A+ Net+
Intel Core i7-950 Intel DX58SO Mobo 6GB Corsair XMS3 Tri-Channel BFG Geforce 260 GTX
2x 160GB Seagate HDs RAID 0 2x 500GB WD RE3 HDs RAID 0
Built 40K+ systems and still counting
 
derFunkenstein
Gerbil God
Posts: 25427
Joined: Fri Feb 21, 2003 9:13 pm
Location: Comin' to you directly from the Mothership

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 12:59 pm

EV42TMAN wrote:
anyone that says you can get away with high end non ECC memory doesn't directly work with databases

That's not true. You know, just based on what I do and what I just said. :p

I just don't wield my job like a hammer and slap everyone with it like it's my dick. Mostly because I don't want to be bothered.
I do not understand what I do. For what I want to do I do not do, but what I hate I do.
Twittering away the day at @TVsBen
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 1:16 pm

Evan wrote:
if you're using the system for something serious like a database, I'd say go for the ECC.
Cool, I'll place the order. I'm glad to hear that this specific setup will have real protection. I was concerned that there was some catch and the full benefits wouldn't be realized on this system (because it's essentaillly an enthusiast machine with a quasi-server single-socket motherboard, and doesn't have true server parts throughout).
derFunkenstein wrote:
The memory controller is what determines ECC support, rather than the CPU in this case (since the memory controller isn't on the CPU), so you should be OK to do that without a CPU swap.
Oh right, it's a pre-Nehalem Intel. How quickly I forget.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 1:27 pm

Memory is cheap enough now that I'd go for it, *especially* if you've got another system that can make use of the non-ECC stuff.

derFunkenstein wrote:
As to what you gain, I dunno. It seems to me that ECC isn't worth the cost simply because if you're getting data corruption errors due to memory, you just have bad memory in the first place and replacing it with cheaper unbuffered non-ECC should fix it just as well.

The point of ECC isn't really to compensate for defective memory (though it can do that too if the errors are all single-bit). The main reason for having ECC is to correct for random "flipped bits" which are caused by cosmic rays (or other background radiation) hitting the memory cells. The rate of these errors actually increases (quite a bit...) with altitude because there's less atmosphere (which acts as a shield) between your RAM chips and outer space. So if you are in (say) Denver, ECC is more important than if you live in Chicago!

derFunkenstein wrote:
The second C stands for Code, I thought. Error Correcting Code. :p

That's... *ahem* correct. :wink:

derFunkenstein wrote:
EV42TMAN wrote:
anyone that says you can get away with high end non ECC memory doesn't directly work with databases

That's not true. You know, just based on what I do and what I just said. :p

I just don't wield my job like a hammer and slap everyone with it like it's my dick. Mostly because I don't want to be bothered.

Well, it kind of depends on how critical the stuff in the database is.

In situations like this, *any* blanket statement of the form "you definitely don't need X" or "you definitely do need X" tends to be wrong. The answer is usually "it depends".

But with RAM being so cheap right now, the only real reason to *not* do it (cost) is weaker than it is normally...
Nostalgia isn't what it used to be.
 
Ryu Connor
Global Moderator
Posts: 4369
Joined: Thu Dec 27, 2001 7:00 pm
Location: Marietta, GA
Contact:

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 1:43 pm

This old FAQ from Corsiar makes some claims regarding the average statistics of a soft error:

The RAM Guy wrote:
Q: Come on... cosmic rays? Really, how often does this occur?

A: The Ram Guy consulted with some experts on this one. Basically, it's a statistics problem. But, when you do the math, a soft error is likely to occur in a system with 256 Mbytes of memory about every 750 hours! And, the more memory you have, the more frequently soft errors will occur.


Those numbers seem a tad high. One would think my 24GB system would be falling over rapidly with values as reported.
All of my written content here on TR does not represent or reflect the views of my employer or any reasonable human being. All content and actions are my own.
 
Krogoth
Emperor Gerbilius I
Posts: 6049
Joined: Tue Apr 15, 2003 3:20 pm
Location: somewhere on Core Prime
Contact:

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 5:27 pm

Ryu Connor wrote:
This old FAQ from Corsiar makes some claims regarding the average statistics of a soft error:

The RAM Guy wrote:
Q: Come on... cosmic rays? Really, how often does this occur?

A: The Ram Guy consulted with some experts on this one. Basically, it's a statistics problem. But, when you do the math, a soft error is likely to occur in a system with 256 Mbytes of memory about every 750 hours! And, the more memory you have, the more frequently soft errors will occur.


Those numbers seem a tad high. One would think my 24GB system would be falling over rapidly with values as reported.


It is because most of those bit flips go unnoticed until it messes up something critical. This is unacceptable for CAD, engineering, medical and scientific related stuff where a single bit flip could mess-up a critical calculation.

For OP, you need to get a X38/X48 board in order to get ECC support for your Core 2 Quad. FYI, not all X38/X48 boards have ECC support, so research carefully on the motherboard features. Asus seems to be relatively safe beat as they tend to offer ECC support on platforms that offer it.
Gigabyte X670 AORUS-ELITE AX, Raphael 7950X, 2x16GiB of G.Skill TRIDENT DDR5-5600, Sapphire RX 6900XT, Seasonic GX-850 and Fractal Define 7 (W)
Ivy Bridge 3570K, 2x4GiB of G.Skill RIPSAW DDR3-1600, Gigabyte Z77X-UD3H, Corsair CX-750M V2, and PC-7B
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 5:46 pm

Ryu Connor wrote:
This old FAQ from Corsiar makes some claims regarding the average statistics of a soft error:

The RAM Guy wrote:
Q: Come on... cosmic rays? Really, how often does this occur?

A: The Ram Guy consulted with some experts on this one. Basically, it's a statistics problem. But, when you do the math, a soft error is likely to occur in a system with 256 Mbytes of memory about every 750 hours! And, the more memory you have, the more frequently soft errors will occur.

Those numbers seem a tad high. One would think my 24GB system would be falling over rapidly with values as reported.

Those numbers are consistent with what I saw first-hand a number of years ago when I worked at Fermilab, and we were looking at DRAM error rates pretty seriously.

Yes, DRAM tech has improved quite a bit over the years, which tends to drive error rates down. On the other hand, as densities go up, the amount of electrical charge required to distinguish a '0' from a '1' gets smaller, making a cosmic ray "hit" more likely to change the charge in a cell enough to flip a bit. I honestly don't know how much effect (or in which direction) these competing processes have had on real-world error rates.

And no, I wouldn't expect your 24GB system to fall over on a regular basis even if the rates are still roughly 1 error per 750 hours per 256MB. Here's why: In order for the system to "fall over" (i.e. BSOD or application crash), you need to either A) corrupt a data address that containts a pointer or array index, in a way which eventually causes an invalid memory access that is caught by the hardware; or B) corrupt a code address (program instruction), in a way which results in either jumping to an invalid code address, or (indirectly) causes an invalid memory access along the lines of mechanism A. At any given time, only a small percentage of your RAM contents will be vulnerable to these types of errors.

Flipped bits in the rest of your RAM will either be C) completely benign (corrupted data will be overwritten by something else before it is actually used for anything); D) mostly benign (result in a fleeting glitch in audio or video playback, or a one-time application misbehavior that goes away if you attempt the same operation again); or E) potentially harmful but not easily detectable (silently flipped bits here or there in your data).

A and B give people who worry about system uptime nightmares; E is the bugaboo for people who worry about data integrity.
Nostalgia isn't what it used to be.
 
Captain Ned
Global Moderator
Posts: 28704
Joined: Wed Jan 16, 2002 7:00 pm
Location: Vermont, USA

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 5:55 pm

just brew it! wrote:
A and B give people who worry about system uptime nightmares; E is the bugaboo for people who worry about data integrity.

How deep are the tunnels at FermiLab, and how deep do they have to be to block cosmic rays (mainly highly-energetic protons)?
What we have today is way too much pluribus and not enough unum.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Thu Nov 17, 2011 6:40 pm

Captain Ned wrote:
How deep are the tunnels at FermiLab, and how deep do they have to be to block cosmic rays (mainly highly-energetic protons)?

The main accelerator ring tunnels are not very deep (just barely below the surface). The tunnels/halls for the neutrino experiments are deeper (few hundred feet down). The stuff I was dealing with was not underground though; it was part of a supercomputer system used for QCD simulations. We started seriously looking into error rates because we had error *detection* (single-bit parity) but no error *correction*, and soft errors were causing simulation runs to crash out at an unacceptable rate (multiple times/day).

Somewhat surprisingly, most data on soft error rates (circa early 1990s) was classified, and since our project was not classified we were not allowed to see that data. So we did our own tests and analysis. We actually looked at leakage from the accelerator as a possible source (and ruled it out). Ultimately we concluded that the errors were being caused by cosmic rays, and low-level radioisotope contaminants (in the packaging of the RAM chips themselves and the solder used to assemble the circuit boards).

Since retrofitting ECC to the system was not feasible, we adopted a checkpointing system where the state of the simulation was periodically dumped to a RAID-0 array, allowing the simulation to be reloaded from the checkpoint and restarted when memory errors were detected.

The electronics industry has subsequently gotten better at controlling radioisotopes in chip packages and solder, leaving cosmic rays as the dominant source of soft errors.

Edit: I don't know how deep you need to be to block cosmic rays.
Nostalgia isn't what it used to be.
 
Dirge
Gerbil Jedi
Posts: 1620
Joined: Thu Feb 19, 2004 3:08 am

Re: Upgrade DDR2 to unbuffered ECC?

Fri Nov 18, 2011 1:22 am

If you find this sort of stuff interesting, there is an AMD White-Paper regarding the use of ECC memory in embedded applications.

Linky: http://www.amd.com/us/Documents/47644A_ecc_embedded.pdf
FDISK /MBR
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Fri Nov 18, 2011 1:22 am

Krogoth wrote:
For OP, you need to get a X38/X48 board in order to get ECC support for your Core 2 Quad. FYI, not all X38/X48 boards have ECC support, so research carefully on the motherboard features.
There's no guarantee that my research is flawless, but the motherboard manual (and the page I linked to in the OP) claim that ECC is supported. Are you seeing something I'm missing?
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Fri Nov 18, 2011 8:57 am

wibeasley wrote:
There's no guarantee that my research is flawless, but the motherboard manual (and the page I linked to in the OP) claim that ECC is supported. Are you seeing something I'm missing?

The 3200 northbridge supports ECC. I would also be rather surprised if any motherboard marketed as a "server" product (other than ultra-low power Atom-based stuff, but that's another story...) lacks ECC support. It is a feature that enterprise server customers demand, and it would be silly for Asus to leave it out.

Many *desktop* boards indicate that they can run with ECC DIMMs, but they are actually just ignoring the extra parity bits, and not using the ECC capabilities. I think your P5BV-C should be fine though.

I did run across the following interesting tidbit, however: https://bugzilla.redhat.com/show_bug.cgi?id=564274 (The TLDR version: Some Asus server boards have a BIOS bug that results in the ECC controller not getting initialized properly if you enable "Quick Boot" in the BIOS; so if you want to use ECC RAM, make sure that feature is turned off!)
Nostalgia isn't what it used to be.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Fri Nov 18, 2011 9:05 am

Dirge wrote:
If you find this sort of stuff interesting, there is an AMD White-Paper regarding the use of ECC memory in embedded applications.

Linky: http://www.amd.com/us/Documents/47644A_ecc_embedded.pdf

Yes, AMD's continued support of ECC across most of their product line (unlike Intel, who apparently sees it as a way of segmenting their product into "desktop" and "server" classes) is one of the reasons I have tended to stick with AMD over the years. This seems to be changing with the Fusion APUs, however; I don't think the FM1 socket has pins reserved for the ECC bits.
Nostalgia isn't what it used to be.
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 3:18 am

The ECC capability is recognized in CPUZ. But Memtest86+ 4.20 reports, "Chipset: Intel 320/3210 (ECC : Disabled)" Does this mean that it's disabled for testing purposes? Or that it's not working (and won't be hen I boot into the real OS)?

I've Googled for 30 minutes and don't see anything convincing. I swapped th CPUs so it now has the Xeon 3360, but that didn't change anything.

Regardless of Memtest, what's the best way to verify that the chipset is taking advantage of the ECC capability?
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 7:01 am

I'm pretty sure the Memtest86+ display is showing the ECC configuration coming in (i.e., it appears to not be working).

Per my post from about a week ago, did you disable "quick boot" in the BIOS? Also check the BIOS for any other settings related to the memory controller.

I'm not sure how to tell whether ECC is enabled in Windows. In Linux you can search the system logs for lines containing the string "EDAC". I suppose you could try booting a Linux live CD to check it...

Edit: Example of checking for ECC using Linux CLI:
mike@quadrupel:~$ cd /var/log
mike@quadrupel:/var/log$ grep EDAC dmesg
[    9.647486] EDAC MC: Ver: 2.1.0 Oct 11 2011
[    9.669510] EDAC amd64_edac:  Ver: 3.2.0 Oct 11 2011
[    9.677842] EDAC amd64: ECC is enabled by BIOS.
[    9.677878] EDAC MC: F10h CPU detected
[    9.677888] EDAC amd64: f10_probe_valid_hardware() This machine is running with DDR3 memory. This is not currently supported. DCHR0=0x3f58090d DCHR1=0x3f58090d
[    9.677890] EDAC amd64:    Contact 'amd64_edac' module MAINTAINER to help add support.
mike@quadrupel:/var/log$

...which actually told me something I didn't realize until just now: the version of Linux I'm running on this system apparently doesn't know how to log ECC errors if you're using DDR3 RAM! :-? Oh well, at least ECC is enabled...
Nostalgia isn't what it used to be.
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 12:12 pm

That was with Quick Boot enabled and disabled (as well as Memory Remap en/disabled). Sorry JBI- after the effort you spent finding and summarizing that long bugtracker, I shouldn't have forgotten to address it (and thank you) in the last post.

And thanks for the example grep code. I got confused interpreting pages of output from the dmidecode command. The grep code says:
EDAC MC: Ver: 2.1.0
EDAC MC0: Giving out device to 'i3200_edac' 'i3200': DEV 0000:00:00.0
I assume the sparseness isn't good. Does this mean ECC isn't enabled, or just that the Live USB key doesn't bother messing with ECC? (Pages like this suggest it's not the latter).

Before I disabled Quick Boot last night, I got the same symptoms that JBI's link described initially ...relentless console error messages like
"gaia kernel: EDAC MC0: UE page 0x0, offset 0x0, grain 1073741824, row 0, labels ":": i3200 UE". But, they didn't show up in the terminal emulator.

After re-enabling Quick Boot, I went back to run JBI's grep code. The output is unchanged. This with a Ubuntu 11.04 Live USB key last night, and 11.11 this morning. This is with the latest bios (released in 2008). Memtest ran 6 hours last night without an error, albeit with ECC disabled.

Is it time to mess with Asus support?

Edit: Corrected "EDAC MCO" to "EDAC MC0
Last edited by wibeasley on Mon Nov 28, 2011 2:19 pm, edited 1 time in total.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 1:19 pm

There may be some other useful output in the general vicinity of the EDAC lines in the log. Try pulling it up in the file viewer, i.e.:
less dmesg

You can search for the relevant section of the file by typing "/EDAC" (without the quotes) and hitting Enter; arrow keys scroll up/down through the file. The fact that your log output is very different from mine isn't conclusive since you're on an Intel platform and I'm on AMD.

I *think* the fact that you were getting the errors when Quick Boot was enabled implies that ECC is turned on, but I am not sure.
Nostalgia isn't what it used to be.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 1:38 pm

One other thought: On the systems I have access to, there does appear to be a 100% correlation between the presence of the "EDAC MC0: Giving out device to ..." line, and ECC being enabled. Systems running the same kernel that don't have ECC RAM installed have a different line, indicating that ECC is disabled.

So maybe you're good.
Nostalgia isn't what it used to be.
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 2:27 pm

I hope so, but I'm skeptical. Did you omit that line for the first post this morning?

I grepped for several other terms, and nothing popped out to me (like 'ECC', 'ecc', 'EDAC', 'edac', and 'enabled'). If anyone is curious, here's the output of "/EDAC" through the less pager.
[   21.607384] EDAC MC: Ver: 2.1.0
[   21.610147] XGIfb: Framebuffer at 0xf4000000, mapped to 0xffffc90013800000, size 32768k
[   21.610150] XGIfb: MMIO at 0xfbcc0000, mapped to 0xffffc90000c80000, size 256k
[   21.610152] XGIfb: XGIInitNew() ...12345678910111215171818118218318618719202122232425OK
[   21.619475] XGIfb: No or unknown bridge type detected
[   21.619478] XGIfb: Default mode is 800x600x16 (60Hz)
[   21.619499] intel_rng: FWH not detected
[   21.622244] XGIfb: Added MTRRs
[   21.622448] fbcon:  (fb0) is primary device
[   21.624433] XGIfb: var->pixclock=25000, htotal=1056, vtotal=1256
[   21.624436] XGIfb: Change mode to 800x600x16-60Hz
[   21.634996] Console: switching to colour frame buffer device 100x37
[   21.644353] fb0:  frame buffer device, Version 0.8.01
[   21.646513] EDAC MC0: Giving out device to 'i3200_edac' 'i3200': DEV 0000:00:00.0
[   21.689529] leds_ss4200: no LED devices found
[   21.864868] sky2 0000:03:00.0: eth0: enabling interface
[   21.865580] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   21.875503] sky2 0000:02:00.0: eth1: enabling interface
[   21.876861] ADDRCONF(NETDEV_UP): eth1: link is not ready
[   21.932076] init: failsafe main process (2920) killed by TERM signal


I wish this page on the EDAC program had a better explanation of ECC detection, or that "Giving out..." entry.

Thanks for all the help JBI.
Last edited by wibeasley on Mon Nov 28, 2011 2:32 pm, edited 2 times in total.
 
wibeasley
Gerbil Elite
Topic Author
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 2:31 pm

This is the only 2GB DDR2-800 on the supported list. Any reason to think I'd have better luck with this, than the one I got?
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Upgrade DDR2 to unbuffered ECC?

Mon Nov 28, 2011 2:58 pm

wibeasley wrote:
I hope so, but I'm skeptical. Did you omit that line for the first post this morning?

D'oh, my bad; that'll teach me to over-generalize. That system is the one exception, due to the version of EDAC bundled with Ubuntu 10.04 not supporting AMD's DDR3 controller. But in that case it clearly indicates that the BIOS has enabled ECC...

On the DDR2 systems I've looked at, there does indeed appear to be 100% correlation. Furthermore, on systems where the memory controller does not support ECC, there's nothing in the log about EDAC at all; and if the memory controller supports it but ECC RAM is not installed/enabled, there's a line that specifically indicates that it is disabled.
Nostalgia isn't what it used to be.

Who is online

Users browsing this forum: No registered users and 1 guest
GZIP: On