Personal computing discussed

Moderators: renee, Flying Fox, Thresher

 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Thu Apr 19, 2012 9:44 pm

I got new RAM for my sig machine today: Corsair Vengeance 2x8GB for a total of 16GB (mobo is mini-ITX has two slots).

I swap it out, run memtest. Lots of Errors.

Go to BIOS. I had set the timings manually: 9-9-9-24. New ram is same speed (DDR3-1600) but rated 10-10-10-27.

Update manual timing settings to 10-10-10-27.

Run memtest. No errors after an hour. Entire pass completes. Looking good. Think I'm going to leave it overnight and it'll still be good when I wake up.

At some short amount of time after that, the screen goes black. I have the monitor connected to the computer via the VGA port on the motherboard: The Intel integrated HD2000. It's not entirely black. There is this slight noise/fuzz at the top left corner. I made a video recording of what it looks like. http://www.youtube.com/watch?v=T8rP03TsXCU

Memtest86 seems to still be running. I cannot see the screen though so I do not know if there are now errors. However, pressing Esc does quit the program: So it did not hang.

You see, what's so frustrating about this setup I've got is that it is so incredibly cramped that the large heatspreaders on this ram actually interferes (1-2mm) with one of the heatsink's fluid pipes. The pipe not being flexible, this requires me to dremel off a bit of the heatspreader.

Now I wouldn't want to do that unless I was sure the RAM is solid and I won't need to RMA it. So I unbolted the HSF radiator so that the sticks could be installed.

And now I get this strangeness in memtest86. :x

Has anybody ever seen this happen?

What does this mean? I'm going to boot into windows and I guess I will run Prime95 on Blend for a while and post back with results.

Update: Prime95 Blend seemed to have caused a reboot about 10 minutes in. I had my back turned to the computer. It reboots so damn fast. This sucks. :cry: Maybe I could find the command rate and set that to 2T and give it a try...
 
derFunkenstein
Gerbil God
Posts: 25427
Joined: Fri Feb 21, 2003 9:13 pm
Location: Comin' to you directly from the Mothership

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Thu Apr 19, 2012 10:30 pm

Mixing RAM from different vendors and especially when it's rated differently can be a real problem. Have you tried the new RAM alone for a while? It may not be bad, it may just be an issue with your other RAM. You'd ahve been better off buying a set that matched what you already have. And if it's not going to fit to the point of cutting the heat spreaders, maybe you're better off buying RAM that will fit. You're going to kill your warranty if you keep it and do that.
I do not understand what I do. For what I want to do I do not do, but what I hate I do.
Twittering away the day at @TVsBen
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Thu Apr 19, 2012 10:38 pm

Yeah. Like I said though, only two slots in this mobo. So this is a new set of two sticks that came together. I've already got the original 2x4GB set in a second machine, that box has 12GB in it now.

I have high hopes for better stability running at 2T (called 2N in the BIOS for whatever reason).

Still not sure what to do about it physically not fitting however.
 
TwistedKestrel
Gerbil Elite
Posts: 686
Joined: Mon Jan 06, 2003 4:29 pm

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Thu Apr 19, 2012 11:42 pm

Even if it's not physically contacting, I'd think the close quarters of everything could be possibly causing a hot spot on one of the DIMMS. If you're planning to use your computer like a normal human, you could probably ignore it and it would never come up - you won't normally be thrashing 16GB of RAM very hard!
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Fri Apr 20, 2012 7:19 pm

TwistedKestrel wrote:
Even if it's not physically contacting, I'd think the close quarters of everything could be possibly causing a hot spot on one of the DIMMS. If you're planning to use your computer like a normal human, you could probably ignore it and it would never come up - you won't normally be thrashing 16GB of RAM very hard!


I fully intend to thrash it as hard as I can -- I am not actually your normal human. There's not likely to be much of a hot spot: the 120mm fan pulling from the radiator goes right over the DIMMs.

So I tested overnight at 2T command rate and still ended up with the same fuzzy screen business.

This morning, rather than testing prime95, I switched to DDR3-1333 speed while leaving it at 10-10-10-27 2T and it has at this point gone 12 hours (almost 8 passes) in memtest and no fuzzy black screen yet. So that confirms this ram is not stable at rated timings (I have it set to 1.575V it is rated 1.50V also...)

And just now as I'm sitting here, 12 and a half hours into memtest, the screen went black-with-fuzz once again.

Boy am I glad I didn't take a Dremel to these. I've got no problem having to run at a lower DDR3 speed but my RAM has to be stable. I'm considering buying server hardware to run ECC in my next rig.
 
The Egg
Minister of Gerbil Affairs
Posts: 2938
Joined: Sun Apr 06, 2008 4:46 pm

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Thu May 17, 2012 11:38 pm

I would be giving the powersupply dirty looks. Do you have a suitable spare to test?
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 5:41 am

I'll be running memtest again now that I've got a proper GPU installed in the box. Hopefully the issue goes away.

I've been able to play plenty of games just fine on the machine: I don't think the PSU is the culprit here since it has no problems handling a much higher power load.
 
NewfieBullet
Gerbil
Posts: 42
Joined: Sat Feb 05, 2005 12:44 am

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 8:54 am

APWNH wrote:
I'll be running memtest again now that I've got a proper GPU installed in the box. Hopefully the issue goes away.

I've been able to play plenty of games just fine on the machine: I don't think the PSU is the culprit here since it has no problems handling a much higher power load.


I very much doubt that this problem had anything to do with the GPU. In fact when trouble shooting this type of problem I would remove the discrete GPU and set everything to stock settings. If you're overclocking the CPU, set it to stock. I am curious what you meant when you said that you had to unbolt the HSF to install the RAM. Is it possible that the HSF is not properly seated on the CPU?
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 12:33 pm

Ah I wasn't being clear. I had to unbolt the radiator. The block has been sitting put on the CPU.

I have decided to clock my 2500K at 4.0Ghz max turbo, stock voltage. It won't do any throttling and should be rock stable.
 
ShadowEyez
Gerbil XP
Posts: 348
Joined: Wed Dec 03, 2003 12:31 pm

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 2:13 pm

Simple testing procedures - try the exact config your box was in prior to putting in the new RAM. Put in the old ram with the same CPU/GPU and software settings, then run memtest and prime95 - does that produce stable results for hours on end?

If so, put in the new RAM at stock timings, changing nothing else. Then run memtest and prime95 for hours. Still stable? Isolate one variable at a time, then re-run tests. Find what change makes it unstable.
The finest tools are forged from the hottest fires
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 2:35 pm

You're right. But there's no way in hell I'm swapping the old ram back in though, I have to unbolt the Hyper 212+ HSF on the other box in order to access the innermost DIMMs. It took me an hour of fiddling to get the damn thing back on last time, with the silly pin you gotta get into the hole and two different possible orientations you can stick the bracket in there. I will not have a repeat of this.

I had done a run at stock 3.3Ghz and it was stable for 24 hours. When I tested 4.4Ghz down thru 4.2Ghz it began to get weird up to ~20 hours into the memtest run. :-? It would seem that I should want to run it for >2 days to ensure any kind of statistical confidence but I don't really have time to do something like that...

At the rate with which memtest performs memory operations, and in a timeframe of 48+ hours, I think we may be reaching expectable memory error rates, considering I'm not dealing with ECC RAM here. Assuming 1 million memory operations per second, at 48 hours that's 172.8 billion ops. Is it reasonable to expect RAM to have a bit of error over several hundred billion operations?

Of course if it was just a regular RAM error it would show up as a red line in the output in memtest, not cause the display to go black.

I'm just running 4.0 at stock voltage now. I think I just don't give a crap anymore at this point. I'll just have it run prime95 in the background for a few days and see how that goes. I need to start getting actual work done on this machine.
 
NewfieBullet
Gerbil
Posts: 42
Joined: Sat Feb 05, 2005 12:44 am

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sat May 19, 2012 3:02 pm

APWNH wrote:
I had done a run at stock 3.3Ghz and it was stable for 24 hours. When I tested 4.4Ghz down thru 4.2Ghz it began to get weird up to ~20 hours into the memtest run. :-? It would seem that I should want to run it for >2 days to ensure any kind of statistical confidence but I don't really have time to do something like that...


If it's stable at stock CPU speed and fails when the CPU is overclocked it sure sounds to me that the problem is the overclock not the RAM. That would certainly suck but I don't think a 2500K is guaranteed to be stable when overclocked.
 
Pheran
Gerbil In Training
Posts: 2
Joined: Fri Jul 27, 2012 10:06 pm
Location: Pittsburgh, PA

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Fri Jul 27, 2012 10:30 pm

OK, I know I'm bumping a 2-month old thread, but did you ever solve this? Believe it or not, I'm having a nearly identical problem with exactly the same RAM - Corsair Vengeance 2x8G DDR3 1600 CAS 10. My whitebox ESXi server is a Gigabyte Z77 board with a Pentium G620 (Sandy Bridge). It was running 8G (2x4G) Corsair Vengeance with no problem. Tonight I swapped that out for the new 2x8G DIMMs and ran memtest86+ 4.20 as a verification. After a while, the screen blanks except for a few moving dots down the extreme left side of the screen. As you noted, memtest is still running because the system will reboot if I press ESC, but pressing other keys will not bring back the screen.

EDIT: Forgot to mention, nothing is overclocked.
 
APWNH
Grand Gerbil Poohbah
Topic Author
Posts: 3200
Joined: Thu Oct 31, 2002 9:10 pm
Location: Boston MA
Contact:

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Fri Jul 27, 2012 11:24 pm

Damn that is really interesting. My board's a Z68.

What I ended up doing was returning the ram. I basically decided it was defective and got a refund for it from Newegg. I'm running a different set of 8GB DIMMs now and they have served me well, no fuzzy screen issue with Memtest.
 
Pheran
Gerbil In Training
Posts: 2
Joined: Fri Jul 27, 2012 10:06 pm
Location: Pittsburgh, PA

Re: Memtest86 4.0 and 4.20 - Fuzzy screen after ~1.5hr test

Sun Jul 29, 2012 10:46 am

Tried testing again with HCI MemTest Deluxe in case this was some kind of MemTest86+ bug, but it seemed to have a similar issue. I'm now testing each stick individually and will report back the results.

Who is online

Users browsing this forum: No registered users and 1 guest
GZIP: On