memtest86+ question with new board and memory...

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Wayne Redpath
Posts: 104
Joined: Wed Jul 29, 2009 7:42 am
Location: Calgary, Alberta, Canada

Post by Wayne Redpath » Tue Oct 06, 2009 10:09 pm

What about those Sharks ,EH...

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Tue Oct 06, 2009 10:37 pm

Wayne Redpath wrote:What about those Sharks ,EH...
I don' git it?

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Tue Oct 06, 2009 10:53 pm

Ok, I ran memtest86+ with both sticks in the 1st two slots (Channels A/B), and got an error. Although this time - though in a similar place in the word (most significant nibble), but a different memory location. No *obvious* connection between the addresses:

When error happened with memory in dual-channel mode (same channel/same colored slots, address was:

12EEE6AC

When error happened with memory in different colored slots, address was:

977736C

Anybody see a relation between those two that knowledge of how memory controllers work might give a hint to?

- Tim

Ksanderash
Posts: 353
Joined: Sat Oct 18, 2008 6:30 am
Location: Moldova, exUSSR

Post by Ksanderash » Wed Oct 07, 2009 3:15 am

tbessie
Have you tried an older Memtest+ release? :) Maybe something is wrong right with the v4.0? And how about Prime95/Linpack?

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Wed Oct 07, 2009 9:41 pm

Ksanderash wrote:tbessie
Have you tried an older Memtest+ release? :) Maybe something is wrong right with the v4.0? And how about Prime95/Linpack?
I had been running an older release (2.something) that came with Ubuntu Server. Same results.

Haven't tried Prime95 or Linpack. Going to try windiag.
Do Prime95 or Linpack (or Orthos, for that matter) run in a non-Windows environment? I'd guess Linpack would, given the name. :-)

Anyway, all I really want is basic memory testing, not system stress testing, at the moment.

But since I saw results of this error (bad particular bit within a byte) show up after some file copies, I suspect it is a true error.

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Wed Oct 07, 2009 10:20 pm

cmthomson wrote:The reason I suggested Ballistix is simple: they are bin selected for high margin to support outrageous overclocks, and they are cheap. (They are also really gaudy; if that bothers you, put some tape over the LEDs :) )

Many other memories are also binned, so don't take this as anything other than a happy customer saying he's happy with a product.


On a completely different tack, if your BIOS will let you bump the north bridge/MCH voltage by a few percent, you should give that a try. This is a well-tested method used by overclockers to make otherwise marginal systems stable. And, just in case it might be a heat issue, make sure your MCH isn't too hot; you might need to upgrade the TIM or airflow on its heat sink...


BTW, memtest is not the most strenuous program for memory stability. I use it as a quick gateway, before running the real bear: ORTHOS. If your system can run ORTHOS for 24 hours in blend mode, then it is rock solid.
I'm just looking for some basic tests, at first. Since memtest86+ is failing (and, I found, Windiag is also failing in exactly the same way, same mem loc, same bit, etc.), I'd like to try some of your MCH voltage advice - but, unfortunately, this motherboard ONLY allows modifications to timing, memory voltage, and "reference frequency" (again, not sure what this means). Given that, would bumping up memory voltage to 2 or more volts be of any use here, do you think?

I'm wondering if Microsoft's own memory test failing just like memtest86+'s will be enough evidence for Intel to let me RMA the board. If not that, then using Crucial recommended memory for the board, and THAT failing, might. Intel is being annoying.

- Tim

Klusu
Posts: 196
Joined: Thu Jul 24, 2008 10:57 am
Location: Riga

Post by Klusu » Wed Oct 07, 2009 10:58 pm

The prime suspect is still the memory. You must get the error with other sticks before changing the MB.
You still haven't tried lower memory frequency, have you?
I have two CM2X4096-6400C5, 1.90V ver4.2. Voltage above 1.8V does not improve anything (other brands work better at higher voltage).

RoGuE
Posts: 540
Joined: Mon Jul 06, 2009 9:11 am
Location: Massachusetts, USA
Contact:

Post by RoGuE » Wed Oct 07, 2009 11:10 pm

I used to own this same memory.

I had very VERY similar problems with it. Memory is pretty cheap these days, so I switched to G.Skill. Haven't had a problem since. *knocks on wood*

For memory to be functioning properly, it needs to pass memtest at stock clock speed, and stock timings as specified by the manufacturer. My corsair memory simply didn't comply. Corsair lost a customer.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Wed Oct 07, 2009 11:26 pm

RoGuE wrote:I used to own this same memory.

I had very VERY similar problems with it. Memory is pretty cheap these days, so I switched to G.Skill. Haven't had a problem since. *knocks on wood*

For memory to be functioning properly, it needs to pass memtest at stock clock speed, and stock timings as specified by the manufacturer. My corsair memory simply didn't comply. Corsair lost a customer.
Don't you think though that the mem loc in error would change if I swapped DIMM slots (with the same mem)?

Anyway, I'm awaiting some Crucial memory that they recommend for this board. We'll see what happens then.

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Wed Oct 07, 2009 11:28 pm

Klusu wrote:The prime suspect is still the memory. You must get the error with other sticks before changing the MB.
You still haven't tried lower memory frequency, have you?
I have two CM2X4096-6400C5, 1.90V ver4.2. Voltage above 1.8V does not improve anything (other brands work better at higher voltage).
Do you mean the timings? eg. stock timings are 5-5-5-18 for this. Should I change that? Or do you mean lower the memory frequency itself, from 800 down to 667 or whatever?

- Tim

RoGuE
Posts: 540
Joined: Mon Jul 06, 2009 9:11 am
Location: Massachusetts, USA
Contact:

Post by RoGuE » Thu Oct 08, 2009 12:01 am

tbessie wrote:
Klusu wrote:The prime suspect is still the memory. You must get the error with other sticks before changing the MB.
You still haven't tried lower memory frequency, have you?
I have two CM2X4096-6400C5, 1.90V ver4.2. Voltage above 1.8V does not improve anything (other brands work better at higher voltage).
Do you mean the timings? eg. stock timings are 5-5-5-18 for this. Should I change that? Or do you mean lower the memory frequency itself, from 800 down to 667 or whatever?

- Tim
no no no...

if the memory you have isn't performing to spec, try some other memory if you have any laying around. If you get the same errors with the other memory, then chances are it's your board.

I still think its your corsair memory tho.

Klusu
Posts: 196
Joined: Thu Jul 24, 2008 10:57 am
Location: Riga

Post by Klusu » Thu Oct 08, 2009 12:20 am

I mean lower than 800.
667, whatever.
Of course, memory must be good at its rated speed. Sometimes it isn't.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Thu Oct 08, 2009 8:37 am

Klusu wrote:I mean lower than 800.
667, whatever.
Of course, memory must be good at its rated speed. Sometimes it isn't.
Ok, just changed mem frequency to 667, running the test while I'm at work, and will report back.

- Tim

cmthomson
Posts: 1266
Joined: Sun Oct 09, 2005 8:35 am
Location: Pleasanton, CA

Post by cmthomson » Thu Oct 08, 2009 2:22 pm

Too bad you can't change the MCH voltage; there's a good chance that would fix your problem.

Some dubious advice above; the fact that the problem does not move when you swap the sticks points at the board, not the memory.

Still, trying different (possibly more tolerant) memory might mask the problem.

Also, changing the memory bus frequency (which I guess you're currently doing) might also mask the problem.

Wayne Redpath
Posts: 104
Joined: Wed Jul 29, 2009 7:42 am
Location: Calgary, Alberta, Canada

Post by Wayne Redpath » Thu Oct 08, 2009 4:06 pm

I have seen no conclusive evidence whatsoever that there is a problem with the Mainboard.

It's very clear that:

= Intel did not claim that those memory modules are compatible with that Mainboard.
= Corsair did not claim that those memory modules are compatible with that Mainboard.
= extensive testing indicates that those memory modules are not compatible with that Mainboard.

My recommendation is to use only certified memory for a new build.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Thu Oct 08, 2009 4:15 pm

Wayne Redpath wrote:I have seen no conclusive evidence whatsoever that there is a problem with the Mainboard.

It's very clear that:

= Intel did not claim that those memory modules are compatible with that Mainboard.
= Corsair did not claim that those memory modules are compatible with that Mainboard.
= extensive testing indicates that those memory modules are not compatible with that Mainboard.

My recommendation is to use only certified memory for a new build.
Whatever the case is, the memory they list as acceptible/tested on the Intel site for that board is very paltry - you'd think they'd mention some of the more known memory brands.

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Thu Oct 08, 2009 4:25 pm

If it turns out I need a new board, and Intel won't send me another one, I'm thinking of getting this:

http://www.newegg.com/Product/Product.a ... 6813128358

Rather overkill, but I like the 8 onboard SATA connectors, and it's gotten rave reviews on newegg, which is rare with that bunch of complainers. :-)

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Thu Oct 08, 2009 10:36 pm

Allright, ran the test at memory speed of 667 (or whatever the one down from 800) is, all day, and it got 10 errors at the same spot out of 16 runs.

It don't bode well for this board.

If anyone can make a good case for the memory being at fault, given how I've swapped the memory sticks between DIMM sockets yet seen the error at the same address, I'd love to hear it.

Tomorrow I get the new Crucial-recommended sticks. And today I bought the above Gigabyte motherboard.

Should I even try the new memory sticks in the old board, do you think? Might they be damaged if the address lines are at fault?

And will I EVER be able to get Intel to RMA the board? The support guy said that typically, I would send them the board, they would run it on a machine on which Windows had been installed, see that Windows would boot, and return the board to me as OK.

That is a really stupid test for a hardware company to use to decide if something were RMA-able or not.

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Thu Oct 08, 2009 11:12 pm

By the way, after I bought this new board, I checked, and the Celeron e3200 isn't listed on it's list of supported CPUs. Just about every other Socket 775 CPU is, but not that one.

Do you think it's worth trying even if it's not on their list?

I've emailed Gigabyte to check, hoping I won't get back a drone-ish "If it's not on the list, it won't work" answer, but something thoughtful.

- Tim

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Fri Oct 09, 2009 6:47 am

Imo you should test the new memory on the old board. Otherwise you (and us) will never know for sure it's the board.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Fri Oct 09, 2009 10:56 am

lm wrote:Imo you should test the new memory on the old board. Otherwise you (and us) will never know for sure it's the board.
I will indeed do that, because I *am* curious. I just hope this board doesn't somehow damage the new memory, if that's possible.

- Tim

Trav1s
Posts: 259
Joined: Sun Nov 09, 2008 12:33 pm
Location: CR, IA

Post by Trav1s » Fri Oct 09, 2009 3:24 pm

tbessie wrote:By the way, after I bought this new board, I checked, and the Celeron e3200 isn't listed on it's list of supported CPUs. Just about every other Socket 775 CPU is, but not that one.

Do you think it's worth trying even if it's not on their list?

I've emailed Gigabyte to check, hoping I won't get back a drone-ish "If it's not on the list, it won't work" answer, but something thoughtful.

- Tim
I would try it...

cmthomson
Posts: 1266
Joined: Sun Oct 09, 2005 8:35 am
Location: Pleasanton, CA

Post by cmthomson » Fri Oct 09, 2009 6:05 pm

tbessie wrote:I will indeed do that, because I *am* curious. I just hope this board doesn't somehow damage the new memory, if that's possible.
No it won't damage the memory, since you aren't running the memory at high voltage. Some motherboards boost the DRAM voltage above what the BIOS shows, but never enough to cause damage when the setting is 1.8V.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Sat Oct 10, 2009 12:14 am

cmthomson wrote:
tbessie wrote:I will indeed do that, because I *am* curious. I just hope this board doesn't somehow damage the new memory, if that's possible.
No it won't damage the memory, since you aren't running the memory at high voltage. Some motherboards boost the DRAM voltage above what the BIOS shows, but never enough to cause damage when the setting is 1.8V.
Thanks for the information. I've got my "recommmended" crucial memory for this board, and am running the tests again.

If it gets the same results, I'll use the new board. Luckily, the new board is spec'd to be okay with the memory I've already got. Not the CPU though, but I'll risk it, as any problems should be obvious quite immediately.

- Tim

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Sat Oct 10, 2009 1:07 pm

Ok, I ran memtest86+ all night (12 passes) on the new memory - and DID NOT GET ANY ERRORS.

That's so strange -- to me (and several other people here) with the tests I did, it pointed to the board more than the memory. So if it was, indeed, the memory/board combination that was at fault, could it just happen to be the an electrical situation is set up using the memory I had in before that causes this occasional error on this one bit on this one byte, even though it doesn't appear associated with one particular stick - just the use of these two in combination?

I'm now running the windiag diagnostic, and will run some more serious ones on it later.

As you can imagine, even though this is now working, I'm still kind of suspicious. I have this other motherboard I bought (the Gigabyte EP45-UD3P) which is beefier and got a lot of good reviews, which I am tempted to use anyway just for fun; or I could return it to frys for my $134.99+tax back, which would be nice as well.

Decisions decisions!

Again I thank all of you for your suggestions and help! Do any of you have an idea why using that Corsair memory would cause errors in the same place, no matter in which slots each stick resided?

- Tim

Wayne Redpath
Posts: 104
Joined: Wed Jul 29, 2009 7:42 am
Location: Calgary, Alberta, Canada

Post by Wayne Redpath » Sat Oct 10, 2009 9:16 pm

re: "Decisions decisions!"

When doing a repair or upgrade on older hardware 'you-gotta-do-what-you gotta-do'. When using current hardware you have the luxury of using qualified memory and CPU. When your hardware works 100% at stock speed then you can use the over clocking features to ensure that you have some 'head room' and then reset to stock speed and enjoy a reliable system for many years.

tbessie
Posts: 232
Joined: Tue Feb 05, 2008 10:07 pm
Location: San Francisco, CA

Post by tbessie » Sun Oct 11, 2009 12:34 am

Wayne Redpath wrote:re: "Decisions decisions!"

When doing a repair or upgrade on older hardware 'you-gotta-do-what-you gotta-do'. When using current hardware you have the luxury of using qualified memory and CPU. When your hardware works 100% at stock speed then you can use the over clocking features to ensure that you have some 'head room' and then reset to stock speed and enjoy a reliable system for many years.
So... not quite sure how you meant that. Would you suggest I return the other board and use the one I have, because the memory I just got for it now works? Or use the new board (assuming the new memory works for it to, which it is rated to do on Crucial's site) as it seems more open to a wider range of memory types?

That is - would you use a board that didn't work with memory X, but works with memory Y, or use a board that works with memory X and Y?

- Tim

Wayne Redpath
Posts: 104
Joined: Wed Jul 29, 2009 7:42 am
Location: Calgary, Alberta, Canada

Post by Wayne Redpath » Sun Oct 11, 2009 6:47 am

I won't tell you what to do. I will educate and recommend and help to clarify. Things to consider are that the Intel board is certified for the E3200 CPU and what are your return privileges, shipping costs and restocking fees.

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Sun Oct 11, 2009 9:20 am

Well, I'd just keep the old one.

Benefits:

- Don't need to reinstall windows, which you might have to do with a new mobo. I'm not sure on this.

- Don't have to take the old mobo out of the case and put the new one in, which is a lot of work.

- The old mobo likely uses considerably less power. See viewtopic.php?p=295510&highlight=x38+p45+x58#295510 for reference.


And some memory-mobo combos just aren't compatible. So the other one is less picky, but basically you found the kind of memory that your old one accepts so there shouldn't be any problem.

Functionally, you shouldn't be able to notice which board is in the PC during everyday use, anyway.

Klusu
Posts: 196
Joined: Thu Jul 24, 2008 10:57 am
Location: Riga

Post by Klusu » Sun Oct 11, 2009 11:09 am

tbessie wrote: ..a board that works with memory X and Y...
You don't know that. Do you?

Post Reply