Asus motherboards + G.Skill memory: Cold boot issues and possible workaround

[WIGILOCO]

I have experienced same sort of issues on boot/windows startup that I get a BSOD.

I have ASUS M4A78T-E and F3-12800-4GBNQ.

I am using BIOS 2303 and 1600MHz 9-9-9-24-2T-1.6V

Today when I go home I'll put them to 8-8-8-21-2T-1.6V on 1333MHz.

[flexserve]

Excellent post!

G-Skill needs to step up to the plate!

I cannot understand how their website claims compatibility yet there are all these issues?

#3 looks good to me..hey G-skill..make mine the Kingston RAM thanks. I am awaiting an RMA in my PM box for the exchange.

Class action lawsuit looks to be forming here..

[Invisible Hand]

Great Post!!!

I just got my Asus Maximus III back from RMA and I'm still having the same Cold Boot problems that I've had for 4 months...I'm pulling my hair out.

Not sure If I should get a new mother board or new ram. But the combo of Asus and G.Skill is terrible.

[cbulman]

I've also been talking about this problem in the Asus forums, and got an interesting reply here from a moderator:

=====

"Thanks for your post... a very complicated problem for sure. What, in terms of memory, could lead to the cold boot problem..? Generally, all these sorts of problem are rooted in the system BIOS... Asus in many cases, writes BIOS code that overrides SPD programming in a memory module .... and so one possibility is that on a cold boot, some program bit in the memory module SPD program is overwritted by the system BIOS and a boot failure occurs... unfortunately, since there are at least 128 programming bits available, most of which are not adjustable in the system BIOS, there is no way to permanently set that bit....

Once you start making adjustments to the BIOS settings for the memory, the system BIOS accomodates... at least for some time.. and finds a working set of data for the memory... but when the system is shut down, those accomodations are lost.

So I wouldn't think that G.Skill is at the heart of this problem.. as you mention, they assemble quality memory. It is much more likely that the problem lies within the Asus BIOS program in relation to the specific electrical characterists of the G.Skill module memory chip revisions."

======

If what he says about the BIOS is true, then I am hoping someone from Asus will post in the thread to let us know that they recognize this issue and are working on a BIOS update to resolve it.

However, I still feel a substantial portion of the blame lies with G.Skill, since they (not Asus) are the ones claiming hardware compatibility. I'd love to hear from either company that they're working together to resolve this, since it doesn't make either one of them look good.

[Swbogie]

Frusterated with GSkill Ram

Good post. I have been experiencing the same cold boot issues for three months. The worst part of this is, a friend of mine gave me 700.00 to build him a computer a couple months ago, and I keep having to tell him that I can't figure out what's wrong with the ram. I'm about to say bye to Gskill and buy another manufactures. That will cost me 100.00 So frustrating

f3-10666cl8d-4bhhk + Asus p5g43t-m pro

swbogie

[cbulman]

First of all, even though this posting may seem a bit long I hope that someone from G.Skill reads this and replies to address this issue.

Like many others who have built a system using an Asus M4A79XTD EVO motherboard with G.Skill F3-12800CL9D-4GBRL memory, I've been having horrible system instability problems. I've been getting occasional BSODs and freezes in Windows, but the most frustrating problem has definitely been the "cold boot" issue that others have been experiencing. This is the problem where the system encounters various BSODs or endless rebooting loops when you try to boot a system that has been off for a period of, say, 8-10 hours or more. I've been working to resolve this issue for over a month now, and it has been a real nightmare. I have probably run close to two dozen different diagnostic tests of every flavor, and have done about a half dozen OS reinstalls trying to get at a solution. It was only by looking online that I was able to identify the culprit, by finding many others who were having the same problems with certain Asus mobo + G.Skill memory configs.

The cold boot issue seems to be a very tricky one to resolve. I have encountered the problem even when using the exact BIOS settings that G.Skill suggests for this memory. I have been in touch with G.Skill support by email and have read their suggestions on this forum and what I have gotten is either reaffirmation on the specified settings for this RAM, or vague offerings on a few other things to try like underclocking the RAM or even messing around with other mobo voltages and frequencies that aren't directly related to RAM settings (not a good sign). I don't think G.Skill has presented a clear answer on this. I have visited just about every other reputable tech forum out there and you'll generally find people out there having our issue, with no consensus on how to resolve it, anywhere.

I want to share what I've learned through all the time and effort I've put into this. Most seem to be approaching this as a problem with RAM timings, which I don't think is the case. Okay, it may be for those using blatantly wrong settings, but I can say that those of us using the spec timings or even looser timings are still having trouble. What I believe to be true is that changing the RAM settings in the BIOS to just about any workable config will correct the cold boot issue, but only in the short term. For example, I was using the "proper" settings for this RAM (9-9-9-24 2T 1600 1.5V) for a while and things seemed just fine, then inevitably the cold boot issue came knocking at my door after a few days. I then tried a different config G.Skill mentioned somewhere (8-8-8-24 2T 1333 1.5V) and lo and behold, stability returned...for a while. Once the badness returned, I was able to restore stability once again by simply switching things back to the way they were before. I think part of why this issue is so frustrating for people is that you can't see whether it's been corrected without having to wait for a while (i.e., leave the system powered down for a while). Also, many think they have the problem corrected by some BIOS adjustment, only to have it return again later. So just to reiterate, I believe that this problem is arising due to some phenomena related to the system being powered down for some time, though I don't have the technical background to know what that might be. I think the memory timing is more of a red herring in this scenario, since the issue crops up with a wide variety of RAM settings, including the "correct" one. For now, what I am doing when cold boot problems turn up is to switch between the two timings I mentioned above. In this way I have been able (so far) to make the cold boot issue go away as soon as it arises.

I would point out that what I suggest trying here is merely a workaround, and not a solution. If you do what I've done and restore your stability, you're probably just kicking the can down the road and will see problems arise again soon. Even if you do what I've done you will get inconsistent system performance due to regularly changing your timings in the BIOS (which is something you should almost never have to do anyway). Plus, it makes me very uncomfortable that I still need to experience the cold boot issue to prompt me to implement this workaround. Having your system crash in the middle of a boot is not good and can easily translate into corrupted data. Still, if others are having similar problems I'd like to hear whether you have any luck with this same workaround.

Here are a few other things others with this issue may want to try to help pinpoint the problem. I will try them myself as I continue to encounter problems and will post what I learn in this thread.

1) Run MemTest86+ for an extended period of time for both the case of A) system seems to be running fine, and B) system is having cold boot issues. The BIOS settings should be identical for both tests. Do you get the same result in both cases, or does one result in significantly different numbers of errors reported?

2) Once the cold boot issue crops up for you, make a small trivial change in your RAM timings and see if that temporarily fixes it (e.g., change from 9-9-9-24 to 9-9-9-27). If this resolves the issue, then I think we can confidently say that this is not a RAM timing issue at all but something else.

G.Skill claims compatibility between this RAM and mobo on their website, so I feel they have a responsibility to consumers who trusted this information. Here's what I would like G. Skill to do to address this problem:

1. Provide some sort of explanation of the "cold boot" problem. Please help us understand why these problems are popping up. This seems to be an issue unrelated to RAM timing and more central to whatever it is that changes if the system is left off for a while. Or, if you think I'm wrong, please explain why.

2. Let us know the full, exact BIOS settings that should be used with this hardware configuration (not just RAM settings, but any other settings that should be noted). Heck, even give us a few stable setting configurations for us to try out. You claim compatibility with this motherboard on your website, so obviously you have determined a working configuration. Make this info a sticky on the 'DDR3 BIOS Configuration Guides' page, as it is badly needed there.

3. If you can't give us the information that will make this hardware "play nice" together, offer to exchange our memory with a different type that offers the same level of performance while delivering system stability. If that's not possible, offer to accept our returned memory sticks for a refund.

I think G.Skill makes good memory, but that maybe in this case they gave this config their stamp of approval without enough testing. If they want to keep me as a customer they need to be proactive about backing up their product claims. I'm willing to try to work with this RAM for a little longer, but my patience grows thin when I read on other forums about people correcting this problem instantly simply by trying different memory.