CMZ8GX3M2A1600C9 failing Memtest in dual channel

Good morning everyone. First of all, thank you for reading my thread, it means a lot. I ask you to please read my thread thoroughly so you can get an idea.

NOTHING is overclocked!

Now, onto my problem...

A few months back I shut my computer down one day. Upon boot up, I received a message that read "current overflow detected, system will restart in 15 seconds". I reset the CMOS, and after that, once I powered on my system, a very weird "shortage" occurred.

Once I tried to boot my system up again, it froze at the motherboard logo screen. During this odd "shortage", my P8P67 Pro Rev 3.0 motherboard at the time, and HDD both died in combat. I called Asus and explained to them what happened, and they setup an advanced RMA on my board and had a brand new Rev 3.1 P8P67 Pro board here within a week, same with the Samsung 1TB drive.

I put my computer together again, and ever since then.. problems problems problems. I honestly can't remember if I BSOD'd this much before I rebuilt my system, but I really don't think so.

To explain the situation, my system worked fine idle. I could watch videos, do this, do that.. but once I played an intensive game... boom, BSOD. I have experience in analyzing and debugging crash dumps, so I was able to read & understand then. Most of the causes were a *3b bugcheck that related to dxgmms1.sys (DirectX) or a core Microsoft file. And as you may or may not know, it is very unlikely that Microsoft files are the causes of crashes as they are protected by the System File Checker, and when you're getting Microsoft file / driver culprits, it's likely a hardware issue.

So, with this in mind, I figured the weird "shortage" may have damaged my Power Supply, and this damage is effecting its ability to handle my system under load. So I contacted Corsair just two days ago, and the tech responded asking to make SURE the Power Supply was the cause before I shipped it off, got a new one, and went through the same thing.

Good idea, so the next thing I do is fire up Memtest before I go to bed. I install it on my USB stick with the pre key USB installer, fire it up... and go to bed. Eight hours later I wake up to over 450 errors.

So, I go to submit an RMA for my memory. It asks for the serial #s and what not which I figure are on the sticks themselves. So, I go into my system, take out the sticks and look. When I realize they're not, I put them back in their motherboard slots and fire up my system.

Upon booting into Windows, which booted just fine prior to removing the RAM, I get an instant 3b BSOD bugcheck.. but it shows no culprit (obviously I also couldn't have checked the dump to see what it was if I couldn't get into Windows). So, I figure I'd go to the next best thing... Safe Mode, nope. I go to boot into Safe Mode and it freezes at CLASSPNP.SYS and then BSODs with the same 3b bugcheck. I tried resetting the CMOS, booting with 1 individual stick at a time, nothing worked.

So, I was pretty much forced to reformat Windows. After I had done that, I was installing all of my updates and such, and in the middle of updates, or after I can't remember, I BSOD'd.

Since I BSOD'd during updates, I had to reformat again because things were just a mess. Before installing Windows again, I dropped my memory from 1600 back to auto to see if that would stabilize things.

I reformatted again, and after installing all of my drivers, all of my programs (hadn't had BSODs or anything yet), RIGHT when I went to install the Logitech webcam software, boom, BSOD. I had to remove my Logitech C510 webcam to be able to boot.

After I FINALLY booted after a nonstop string of lvbflt64.sys culprits (Logitech Webcam Software driver), I tried to uninstall the software via control panel, but it said it had already been removed. By that, I can only guess it wasn't fully installed, so there wasn't really any possible way to automatically uninstall it.

So, I manually uninstalled everything, even in the registry (just did a ctrl-f for "Logitech" until it couldn't find any more entries). I also deleted lvbflt64.sys from sys32/drivers manually.

After I did all of this, I restarted, plugged in my webcam... didn't BSOD this time, and forced Windows to manually install drivers for it. After that, my webcam was recognized and I was able to use the microphone on it as well. This is weird to me because I have never had an issue installing the webcam software before, but I am not going to try and install it again right now.

So, as I said, after dropping my memory from 1600 to Auto, I played on my PC all night with a few friends in a very non-intensive game. After we all got off, I fired up Memtest with both sticks inserted at Auto speeds rather than 1600 this time to see if I got no errors.

I went downstairs to do a few things and came back up not even ~15 minutes later to find two red errors waiting for me.

So, what are your suggestions? Does this sound like some tuning needs to be made in the BIOS for my system, or is my memory bad?

Oh, and just to clarify:

1. I am running both sticks in their recommended slot configurations based off of the manual. To be specific, DIMM slots a2 and b2 (both light blue slots).

2. The timings I am using with both sticks inserted is 9-9-9-24-2t command rate.

3. The DRAM voltage I have set is Auto, which is always hovering at 1.5v, or a little above according to the EFI BIOS.

I would suggest checking for the latest MB BIOS and then go to0 BIOS setup and load setup defaults and enable XMP and disable legacy USB then test the system with http://www.memtest.org and see if you still have errors. If you do get an error I would test the modules one at a time at least three passes.

Thank you very much for your response, Ram Guy. I just have a few questions if you don't mind!

1. You say check for the latest MB BIOS. I assume you mean ensure that I have the latest BIOS version for my motherboard, in which I do. I in fact updated it yesterday before this latest installation of Windows.

2. If I load setup defaults, I believe it sets the timings to 9-9-9-24-1t command rate, rather than the advised 2t. Is this still what you'd like me to do?

3. I was reading a few threads and noticed you mentioned that disabling Legacy USB Support on P67 boards was important, why is this? I'm just curious.

This error is a possible sign of what may be going on. have you seen this error since and I assume that you got this error as a BIOS post message before the system Booted to Windows is that correct?

Also have you tried removing one of the Video cards to see if the system is more stable?

I have not seen this specific error since I had my motherboard replaced, and yes, you're correct. On my old motherboard, when I booted up my system, I believe it showed this message before even displaying the motherboard logo, but I cannot remember. Regardless, yes, it was most definitely before getting a chance to boot to Windows.

And no, I have not tested a single 5850 at a time yet. Would you recommend that before or after testing my memory?

With this being said, what do you think could have caused that error, and what relevance are you thinking it has now?

As soon as I receive an answer towards my questions to make sure I am doing everything correct, I will move forward with my memory testing. Once again, thank you very much!

Wow, you're fast. I cannot thank you enough :)

A: If you are getting errors then yes set it manually to 2Tcommand Rate is not really a memory setting it will set the memory controller more aggressively and as such the memory may have little to do with it.

That's interesting, I did not know that. The only reason I was setting my RAM to 2t command rate was because I thought that was the advised factory command rate.

Alright, I will disable Legacy USB Support!

And to just be sure so we can have flawless testing here, you want me to set the command rate from 2t back to 1t, correct?

After I've done both of these, I will run Memtest again. If I get errors with 1t command rate on both auto freq and 1600 freq, I'll remove a video card and test again to ensure this is not PSU related issue.

Gotcha.

And sorry for so many questions, I just want to have a flawless testing phase here so I can provide the most accurate info for the most accurate support... how do I actually enable XMP on my mobo in the BIOS?

I can never find such an option. All I know is, after I set my memory to 9-9-9-24-2t, auto memory frequency and 1.5v auto dram voltage, in Memtest it says *XMP* next to each stick. Does this mean XMP is enabled?

It will be under the AITweaker menu.You should have three choices, maybe 4. Auto, manual, and XMP. Just choose XMP , save and exit!

n Memtest it says *XMP* next to each stick. Does this mean XMP is enabled?

What frequency is memtest reporting?

Ah, found it!!! Thank you so much, Peanutz!

Alright, I have set it from Auto to X.M.P and have disabled Legacy USB Support. Before I go to bed, I'll run a Memtest and wake up to the results.

Question, should I keep Legacy USB Support disabled at all times from now on, or just keep it disabled when running memtest?

Have a good night.

Question, should I keep Legacy USB Support disabled at all times from now on, or just keep it disabled when running memtest?

Just while testing. If left off it may do just what it's name implies....it will disable some older USB devices. :) Keyboards, mice and so on. :)

Well an update before I go to bed:

In the UEFI BIOS, there's two Legacy USB Settings. The first is just titled "Legacy USB Support" I believe, and you have the options of Enable, Disable, and Auto. I disabled it as that is what I was told to do, but upon reboot into the BIOS after a save, it no longer detected my USB drive, so I couldn't boot Memtest.

I changed it from Disable to Auto, which I believe read something along the lines of "Disables Legacy USB Support if no USB devices are connected". After I did that, I was able to boot off of my USB stick again and run Memtest.

The second Legacy USB Support option was titled "Legacy USB 3.0 Support", which I believe only had the options Enable and Disable. I didn't touch that as I didn't know if it was necessary or not.

Regardless, here's what I did so far:

1. Reset the CMOS.

2. After resetting the CMOS, I navigated to AI Tweaker and set it from Auto to X.M.P which then did everything itself.

3. After it set everything up, I saved settings and restarted.

4. Changed Legacy USB Support from Enabled to Auto as Disabled wouldn't allow my USB stick to be detected in the UEFI BIOS.

5. I ran Memtest, and no longer than ~10 minutes, I received over 90 errors.

What would you suggest I do next? Remove my second 5850 and test again, or what else?

Thanks, have a great night. I will check this thread when I wake up tomorrow afternoon (I am nocturnal :P)

And just to make your life a little easier by providing some more info, here's screencaps via CPU-z of my memory settings currently:

Memory tab:

http://i105.photobucket.com/albums/m212/momorox_01/PC/memorytab.png

SPD slot 2:

http://i105.photobucket.com/albums/m212/momorox_01/PC/spdslot2.png

SPD slot 4:

http://i105.photobucket.com/albums/m212/momorox_01/PC/spdslot4.png

Do those look right?

There are three things that come to mind.

First to solve the "Legacy USB' issue, can you burn Memtest to a disc? That would allow you to turn off the USB support to rule out false positives. AS you found out you loose your USB ports.

If you don't have an optical drive , test each module individually with the XMP settings and Command Rate at 2t. That would serve two purposes.

1) it would give us an idea if any one module is to blame

2) If they all fail,or all pass we can be certain that it's some other issue. If just one module fails , then it's pretty cut and dry. Rma the RAM!

Just be sure to load set-up defaults first , then enable XMP before testing.

And of course , like RG suggested test with one GPU. Which may be something you want to try first just because it is a quick test compared to running memtest on all the RAM. Just unplug the power cables.

Do those look right?

Yup, those look just fine!

Thanks for your reply!

I do not have access to any blank CDs at the moment, so I'll have a friend bring one over sometime today. Once I get that disc, I'll do the following:

1. Disable Legacy USB Support

2. Run Memtest again on the CD rather than the USB

If I am still getting errors, I'll test on module at a time. If I am still getting errors, I doubt it's a power issue and we can pretty much conclude it's the RAM, right?

Also, as far as Legacy USB 3.0 Support goes, should I keep that enabled when disabling the other, or disable that too?

My friend didn't have CDs, only DVDs. When I tried to burn it, it kept saying "Invalid image file". I tried the old way of just throwing the files on the disk, right click and then burning to disc, which went through... but when I go to boot from my DVD drive, it just does the flashing _ for a bit, and then goes right to "Starting Windows".

So, for now, Memtest on a disc is out of the question.

What do you suggest I do next?

Here i'll copy this. LOLBeen a long aday ! :)

If you don't have an optical drive , test each module individually with the XMP settings and Command Rate at 2t. That would serve two purposes.

1) it would give us an idea if any one module is to blame

2) If they all fail,or all pass we can be certain that it's some other issue. If just one module fails , then it's pretty cut and dry. Rma the RAM!

 

 

Just be sure to load set-up defaults first , then enable XMP before testing.

 

And of course , like RG suggested test with one GPU. Which may be something you want to try first just because it is a quick test compared to running memtest on all the RAM. Just unplug the power cables.

Let us know how you make out!

Alright guys, good news! At least a break from all of the errors...

Before testing one module at a time, I removed my second 5850. At first, I just removed the pci-e cables attached to it and left the bridge attached (crossfire bridge), but when I turned my system on, my motherboard screamed at me. After I actually took the second 5850 out of the system, I booted up just fine and went to Memtest.

Just before I removed the 5850, I did a test Memtest to see how long it'd take to reproduce errors with the X.M.P profile enabled, it took roughly 3 minutes.

After removing the second 5850, I ran Memtest again and it went strong with no errors for 10 minutes. I stopped it at 10 minutes because I pretty much got my answer that power may just be the issue here.

I of course will run Memtest over night to be 100% sure this is the culprit.

So, now that I've told you this, and Memtest didn't fail around the time it failed with the second 5850 in, that tells us this is a PSU issue?

If it is, Corsair has been wonderful enough to accept my RMA and give me 90 days to send it in, so we have plenty of time here! :)

Thanks guys.

Good afternoon guys, I bring more good news.

As I said, I ran Memtest before I went to bed last night to be sure it was the PSU ever since I removed the 5850. Well, after 5 tests and only a few minutes in, I got errors. So, I began 1 stick at a time!

First off, I put 1 stick in the recommended slot for single channel and tested, failed within minutes. Second, I removed that stick and put the other one in and I woke up to over 8 hours of memtest with ZERO errors!!!

So, this is a single stick issue!!!

What do you guys recommend I do? I lost the box for this RAM, so I believe I'm missing some essential information for the RMA of this RAM :P

**When testing the RAM in single channel, I was using X.M.P. First stick failed, second stick went strong for over 8 hours and was still great when I woke up.

What do you guys recommend I do? I lost the box for this RAM, so I believe I'm missing some essential information for the RMA of this RAM :P

Just use the link on the left to get your RMA started. You dont need the original packaging just the kit number. Just make sure the modules are well protected when you ship them. A padded envelope works great.

Thanks for the reply.

In the RMA, it's asking for the Serial Number and Lot Codes of each module. Where do I locate these? I cannot find them on the sticks, just the model #, some timing info and a bar code.

As I said, I don't have the box anymore so I cannot locate the information on that :(

Edit: Ah, I read right above that "If your module(s) do not have the lot code or serial number, simply enter NA in the field. Providing as much information as possible will help expedite your RMA." :P

Well, I submitted the RMA and chose the "Ram Guy has already helped me method", and linked his first post.. so they can read the thread and see :). Thank you all for your incredible help, Corsair really is a great company with outstanding support.

Just a post before I go to bed because this is starting to make zero sense to me.

So when I was testing the sticks individually, I had the 2nd 5850 removed. As I said, the first stick I tested failed within 3-5 minutes, and the second stick I woke up and was going strong with no errors after 8+ hours. When that single stick passed for 8+ hours, that was with the 2nd 5850 removed. So, after waking up yesterday I plugged the 5850 back in.

I just ran Memtest a few minutes ago on the SAME stick only that passed 8+ hours with NO errors, but this time with the 5850 installed, and I got errors within minutes.

I am so frustrated at this point. I don't know if it's the RAM or if it's the PSU. Regardless, I have both RMAs accepted (PSU and RAM). I might as well just RMA both at this point, because with my luck I will get the memory back and it will be the PSU.