Memtest stops, then passes - is my RAM ok?

[soundasleep]

I have a new 2500k / Gigabyte H67 build that I am still testing, not using it for anything yet. The RAM seems to be passing Memtest (latest version) with no errors, but it never seems to run completely the first time.

First ran it with just Crucial 2x4GB CT2KIT51264BA1339. First try, it stopped without completing the pass. Rebooted and tried again, ran through a few passes with no errors.

Added G.Skill 2x4GB Ripjaws for a total of 4 modules. First try, Memtest stopped after about 5 minutes. Second try, stopped after about 10 minutes. Third try, ran all night (3 passes) - no errors.

The only thing I did during the test was plug/unplug a DVI cable. I'm thinking the RAM's all probably fine, right? I'm just a little concerned since I've read that when Memtest just stops, it's not good.

 

[wirednuts]

i had a huge headache on my hands for almost 2 years until i finally found out my ripjaw ram really was bunk. the reason i never suspected the ram was because it always passed memtest86+ tests. the problem definitely was the ram though, i eventually replaced all the parts in the computer besides the ram, and nothing fixed the problem until i replaced that one of two sticks i had.

 

[soundasleep]

Well, that's a little disconcerting, especially since I didn't really NEED the extra 8GB. I'm thinking I'll run the Windows 7 memory diagnostic tonight. Any other way to find problems?

 

[taltamir]

you write "stops" but from your description it seems to me that the system crashes, hard, and reboots. There is definitely a hardware problem with it just randomly reboots while running memtest. There is a good chance that the problem lies with something other then the ram though. I would guess the problems to be:
1. Mobo
2. PSU
3. CPU
4. RAM

 

[soundasleep]

Sorry I wasn't clearer. Not a system crash/automatic reboot. Memtest just stops running. The time counter stops, no more % increase in the current test. It just stops running. Actually, one time I think there was actually a message, something like "The test has been stopped/aborted" (but no use of the word "fail").

At that point, I couldn't get out of Memtest with ESC or any other key, so I had to hit the reset button. There has been weirdness a few times, when I used the reset button in these instances. The machine would start booting, but then my Asus DVD-RW would make a loud clacking, and the system would automatically restart over and over until I go 4 seconds on the power button. I was assuming this was something with my DVD-RW drive. Ok, maybe it is a crash of some kind... ugh...

 

[taltamir]

interesting. I would ask in the memtest forum.

 

[soundasleep]

Thanks taltamir, I just checked over there and actually found a few people with the same problem (Memtest stopping). They would sometimes get the same message I got: "Testing xxxk - xxxxM ...Halting". Turns out it is probably the USB mouse, apparently a mouse move is being interpreted as an ESC press in some systems. Explains why it happens sometimes but not others. I'm still gonna run Win 7 memory diagnostic but this makes me feel better.

Still doesn't explain why I get that weird reboot cycling when I use the reset button sometimes (and only when I use the reset button...which ideally shouldn't be often of course...).

 

[bryanl]

I would suspect changes in temperature, which can change voltages and the signal voltage threshholds of chips. That's rarely a problem with A-grade, like those on the Crucial memory, but the Ripjaws were made with B-grade, as indicated by the ghetto name.

 

[kmmatney]

I would also try the Microsoft memory diagnostic CD. I've had better luck with that versus Memtest at finding memory problems. Just google Microsoft Memory Diagnostic

 

[soundasleep]

I just tried the Windows 7 Memory Diagnostic Tool (not from a CD, but it did boot into DOS to run it) and there were no errors. I think I'm good. Not sure what exactly was causing Memtest to halt (false ESC command, temperature, ?), but if no errors are being found, I don't know what else I can do...

 

[Patrick Wolf]

Still doesn't explain why I get that weird reboot cycling when I use the reset button sometimes (and only when I use the reset button...which ideally shouldn't be often of course...).

That's most likely the Giga board.
http://www.overclock.net/intel-motherboards/1034724-gigabyte-p67-z68-boot-cycling-problem.html

I had the Z68X-UD4 with the same issue, had to return it once the boot cycle become eternal. Even so, I ran memtest overnight on it which made 9 passes and no errors before I stopped it. I used V4.20 auto-installer for USB Key (Win 9x/2k/xp/7). And my memory was http://www.newegg.com/Product/Produc...82E16820231445.

 

[soundasleep]

Thanks Patrick, that is very helpful. I read the first page of the thread and it sounds like my exact problem. Since I'm not using the computer for anything yet, I don't know if this cycling problem will be happening any time other than when the Restart button is pushed (which causes it pretty much every time).

Now that I think about it, I believe the first time I used the Restart button, the MB (GA-H67A-UD3H-B3) had to boot from the backup BIOS. Since then it just does the endless cycling thing. And I've seen that "Overclocking failed" message a couple times too. ARGH. Well at least it's a known problem.


So my next question...before I dive into all 11 pages of that thread...is this fixable with a BIOS update or something? Or does this mean the MB needs to be RMA'd, no hope for it? Or is this just how this MB is, and I need to live with it? Did you get an exact replacement for your GB board and it worked?

I don't think it's my RAM, as some have theorized in that thread - as this was happening when I was just using approved Crucial RAM.

 

[Patrick Wolf]

Not sure what fixes it, if anything. Some say a bios update fixed it, others not. I also tried different memory with my board and it didn't change a thing. It doesn't really seem to be a memory issue. From a cold boot, my board would usually loop 3+ times before posting and loading windows (where everything was stable). Then after 12 days of that the board refused to even post, it was just an endless loop. I returned it to newegg for refund, they even gave me a prepaid shipping label which is odd cause their policy says customer pays return shipping on defective items. I haven't ordered a replacement, but it will not be another 1155 Giga board.

There are other newegg reviews describing this issue as well. Though most say they have no problems at all so that makes me think the boards with this issue are just bad and may not be reliable long term.

 

[wirednuts]

the only way you can tell for sure is to swap with a known-working stick of ram. and you have to know that stick works with your board, so never rule the stuff out until the problem is found.

 

[soundasleep]

Ok, I just read through all 19 friggin pages of that thread. Very frustrating issue. I can say that the "boot cycling" description that you agreed with, is exactly what happens with my board. I think I'm going to post on that forum since I have an H67 board (i.e. different chipset and no overclocking) and that would be a first for that thread. It would eliminate the theories of some of those guys that it's OC or chipset specific. It seems to be a problem with all SB-related Gigabyte boards, from what I can tell. But apparently not EVERY board. So it sounds like I probably have a bum board.

What a f'n pain. I really don't want to have to RMA this thing. I'm still testing HDDs and stuff so I'm not using the system quite yet. I guess I'll try a few BIOS versions and see if that helps (not expecting much). I suppose I should try the single-memory-stick thing too, just to make sure it's not the RAM. If those don't fix it, then I have to decide if I'm gonna take a wait-n-see approach and see if it gets worse like yours did...or just nip it in the bud and return the thing. I guess I'm leaning toward returning it, since it's probably a sign of things to come. UGH. The idea of my new build was to go for stability and quietness (hence no OCing, no video card, etc.) and I'm already failing. Or Gigabyte is.