Cold Boot Issues - G. Skill PLEASE READ AND RESPOND

[Grimmett]

Hey everyone,
I've been monitoring the forums for a couple weeks now. I'm seeing more and more users experiencing the exact same, or very similar "cold boot" problems. This seems to be happening not only from users using all four DIMM slots, but many users using only 2X2GB (2 DIMM slots filling one mobo channel).

NOTE: I have done all of the tweaking suggested by G. Skill and it hasn't helped.

Basic Symptoms (w/ 2x 2GB filling one motherboard channel):

1. System runs fine and stable while warm. All Memtest and Prime95 tests pass.
2. After the computer is turned off overnight (6-7hrs), the computer is restarted. Windows either BSOD before or around login, and/or Windows error messages popup saying all sorts of services were either terminated or have stopped running.
3. Upon restarting the system, Memtest is ran and 100-1000's of errors appear.
4. One 2GB stick is removed from the channel. The computer is restarted, and Memtest is performed with only one 2GB stick remaining in the channel. Memtest completes many passes with 0 errors.
5. The second 2GB stick is replaced, Memtest and Prime95 passes, and everything is stable until the next time the system is turned off for a period of 6-7 hrs.

I'm just wondering if G. Skill has figured out what is causing this issue for so many people? Was their a manufacturing problem with the RAM? Was there some sort of issue with QC? Or, possibly, is there just something that alot of us are doing wrong? Myself, and I'm sure many others, are curious as to why this is happening with so many different models of RAM?
What is different with the replacement RAM that you are sending people with RMA's? I understand that other memory companies are experiencing these problems and I am not pointing fingers. There just hasn't been any sort of formal response or explanation from G. Skill as to the cause of this problem and the steps you are taking to correct it. Can we be comfortable knowing that the replacement RAM will work correctly? I'm just very uneasy that the replacement RAM will get here sometime next week and it not work.
I have spent over $120 on this RAM and shipping it back to Newegg and G. Skill, and so this whole experience has been very disheartening and costly.
Can you please calm my fears and explain to us about the status of this issue, and what it being done to ensure that we are receiving quality products in return?

Thank You!

[Tiribulus]

This won't be very helpful and we already talked about this, but I still doubt it's actually the memory. I can only be honest.

[Grimmett]

Have you not read the countless other posts from people having the exact same problem? How naive can you be? I can only be honest.

[Tiribulus]

Have you not read the countless other posts from people having the exact same problem? How naive can you be? I can only be honest.

No need to take your frustration out on me friend. I'm not trying to minimize or discount anybody's problems and I have plenty left to learn. I just don't understand by what principle a memory module can stop working and then start again. Maybe they can, but I've never seen it and am only giving an opinion.

[Grimmett]

Wow, I am so sorry. I read that reply early this morning and totally took it in the wrong context. I thought you were saying my post wouldn't be very helpful. I am EXTREMELY sorry. I should have some coffee before reading these things I guess. :P
While I'm not exactly sure what is wrong, I think it is at least related to the memory. One thing that comes to mind that would cause memory to work perfectly normal at one point and haywire at another would be temperature. When the system is cold, including the memory sticks, there could be a possibility that something isn't contacting the way it should. When the system warms up, everything is peachy. I thought of that reading another post where someone had a bent pin on their CPU and was actually getting good contact with a warm CPU and bad contact when it was first booted. I suppose the contacts on the memory sticks could act the same way if they were defective in some way.
Oh well, we'll see when the replacements come. Again, very sorry.

[Tiribulus]

Cewqrz

No trouble. Having reread my post I can see how it could be taken that way, but yes, I was saying my post wouldn't be too helpful.

Unless G.Skill is knowingly passing failed modules I'm not sure what else they can do. I mean put yourself in their position. You spec out a product, test with a bunch of boards and deem them "compatible". You then reproduce that product, but, unless we wanna be paying 500 bucks for a stick of ram there is no way each individual piece can be tested with every certified board so they do a generic functionality test and the ones that pass are sold. What else can they do?

I gotta tell ya man, I could be wrong, but my money is on the motherboards and power supplies in many of these cases even where memory tests are failing. I can't say enough that I do not know everything, but it seems far likelier to me that a warming up effect would be found there than the memory. The real kick in the nads is that somebody could have non faulty parts that still won't play nice together. At least easily.

Years back I had a pair of Abit (yes, I loved Abit) NF7-S's. One for my daughter and one for me. I gave her some middle of the road Samsung sticks I had there and bought myself a pair of better OCZ modules. All were 512's. A while later it became known that a certain batch of Corsair memory was exceptionally overclockable and Newegg had some. I grabbed a pair (they were gone by that afternoon) and put em in my machine. Good stuff. I figured I'd give my daughter my old OCZ memory to replace the Samsung she had in there. Nope. Would not boot fer nuthin. Get this. Her old Samsung memory WORKED in my NF7-S and my new memory worked in hers too. Thinking something might have happened to the first pair of OCZ sticks I put them back in my machine and they worked fine

So I had a pair of Samsung sticks that worked in both boards.

A pair of OCZ sticks that worked in both boards.

AND a pair of OCZ sticks that steadfastly refused to run in one NF7-S, but ran fine in the other. Never did figure that one out either.

Today it's even worse with many more different chipsets, processor families and memory options etc.

This post won't probably be too helpful either. Sorry.

[jsbeddow]

Same cold boot issues here.....

I just wanted to "second" Grimmett's request to GSkill in some transparency on this "cold boot" issue. I too have been battling this problem for about a month, and finally isolated it to the memory (tried swapping out power supplies first, then re-installing Windows 7 64 bit, and trying different drivers, all to no avail). I am currently awaiting my RMA replacement from GSkill on this, and am crossing my fingers that the new ram will work correctly, but I probably won't know for sure for over a week after receiving the new ram, due to the unpredictable nature of the problem, and the need to only test it after being "off" for 10-12 hours or more.
Like so many others, I have no problems getting the system to run stable (Prime95 and Memtest) when "warm". I too had trouble believing it could really be temperature related, especially when the trouble is at the cool end, not the usual problems that crop up when components get too hot. For anyone interested, my board is an MSI P55-GD65, CPU is an i5-750, and the ram was the GSkill F3-12800CL9D-4GBNQ (2 sticks of 2GB each).

Please, can someone from GSkill give us some insight into the cause of this? What has changed that allows them to be so confidant that the replacements will work correctly?

[Tiribulus]

<<< What has changed that allows them to be so confidant that the replacements will work correctly?

If there really is something like this then I agree. It would go well for the company to declare it up front. I, for one, even though I haven't had this issue, would be very interested to learn how a simple memory module could require warming up before functioning properly. It just feels wrong to even type that.

[espac]

I have the same problem. see my thread. I have send the motherboard and the ram to RMA and still cant figure whats happens, the problem is still there. tomorow i will do video and upload to youtube and you will see whats happens.
http://gskill.us/forum/showthread.php?t=5184

[Grimmett]

Update

Check out this link: http://gskill.us/forum/showthread.php?p=21615#post21615

G. Skill has given some insight to the situation. I'm not sure if this "revision" of the RAM will fix my problem or not, but it gives me hope. I wonder what the revision was? We'll see how all the replacement RAM works for everyone, I guess!

Good Luck

[espac]

I said that i have send the RAM to RMA and i recived the new one this monday but the problem is still there. Its true that the blue screen that appears is not the same that witth old ones. this is the blue screen with the old RAM
http://img200.imageshack.us/i/1004472w.jpg/
in the new blue screen it sais" dumping phisical memory 1-100" and then restart the pc.
I dont know what type of revision they have made but the problem is still there

[Grimmett]

Blue screens can happen for alot of different reasons. Your issue may not be because of bad RAM, so hang in there. I know it's frustrating and costly. Have you tried a different power supply, by any chance? You could get the same memory errors if your memory/motherboard isn't getting the voltages it need, I would think.
The problem does sound similar to mine, though. Maybe your replacement RAM wasn't part of the revision? I wonder when the revisions were made? Hopefully G. Skill or someone else can give you some more advice. I hope you don't have to send it back, again... I wish I could have been of more help.

[Tiribulus]

Blue screens can also happen because of not enough voltage to any one of the parts because of being set too low. The readings you're seeing may not even be that accurate anyway. If your bios says 1.5 for the memory it may actually be only 1.47 or something. It may also be a little over as well though. This machine is overclocked quite a bit

My G.Skill memory is rated at 1.5. I have it running at 1.66. I also have the NB/CPU, CPU, NB and HT voltages boosted as well. If I drop any of those I get weird blue screens that I'm certain are telling me nothing useful about the problem. Don't be afraid of a couple tenths of a volt unless your cooling is really inadequate. Especially with system memory. It can take it and trust me, even without heat spreaders will not have a problem with more juice unless it's way way over spec. All spec means is it should be able to run that low. When I get some I'm betting I can get some time more muscle out of this memory and it will wind up over 1.7 for sure at that point.

For testing I would try a pretty fair jump in all your voltages just to see if the problem goes away.

[GSKILL TECH]

espac,

Create a thread and post pictures of your BIOS and we can see what's going on.

Thank you
GSKILL TECH

[Tiribulus]

Something to remember too is that if this is PC3 800/1600 the board is probably expecting 1333 which means even if you set the speeds correctly it may still need juice on the NB/CPU and NB at least, just for the memory to run at it's rated speed.