Random freezing

[magibeg]

Check system for my stats and such. I recently put in a set of G.Skill ram a few weeks ago so they seem like a likely suspect. I can run prime on small FFT's or Blend and never seem to come up with any errors. While in vista the explorer bar will freeze and stop responding, ie essentially dies, and my desktop stop responding. However programs i'm already running such as msn messenger or prime still operate just fine(without errors as everything else goes to the crap).

Tried running the windows memory diagnostic tool and it got to 21% or so then froze, without any errors of course. Things seem to run fine in safe mode however and i don't get any of the freezing. Basically I'm just getting confused with the mixed signals my system seems to be telling me

[sneekypeet]

Get memtest86+ 2.0 or better and burn the ISO to disc. Then memtest will run in DOS and show errors versus failing. http://www.memtest.org/

Edit: in case you dont know...Nero or say magicISO will burn it to a disc for you as well as many others, but those are what I use on a regular basis. http://www.magiciso.com/

[Kursah]

Do what sneeky suggests...maybe downclock the ram and see if it's stable at stock specs (dunno what kind of specs that mem has for stock). It sounds like unstablility to me magibeg...but also remember Small FFT's in tests is CPU Only stuff...it uses so little memory that it's not conclusive enough to even be a partial memory test, Large FFT's though will be decent for testing CPU/NB/MEM/etc.

Keep us posted man! I think you have a voltage setting too low or a timing too tight, just kinda depends on what your it is..I don't really think 5-5-5-15 would be too tight tho. How did you verify stability on that ram when you installed at and set it at that speed for that system oc in the first place? Was it stable then?

[magibeg]

Well i'm running memtest now, and it was stable as stable could be before. I more or less ran small FFT's to eliminate the risk of it being my cpu, i dropped everything down to stock for testing purposes of course. At stock the ram is 1000mhz 5-5-5-15 so i haven't really touched that. Dropped it down to 800mhz later for testing and i still suffered from the same issues. We'll see what memtest turns up over the next several hours, i'm actually hoping for errors because it would give me a solid answer to things.

[Kursah]

Do you have the 2x2GB blue kit? If so, I have always been stable with 5-5-5-15, 2.0v...I am able to get a little OC out of 2v, I wouldn't see running DDR1022 being an issue.

What mem voltage are you using? How much NBv/MCHv have you given? You might just have enough to run now but not enough to be stable...parts do degrade over time, some faster than others unfortunately as you know...so maybe a small bump up in voltages in this area would help.

Keep us posted on your results!

[magibeg]

Originally Posted by Kursah

Do you have the 2x2GB blue kit? If so, I have always been stable with 5-5-5-15, 2.0v...I am able to get a little OC out of 2v, I wouldn't see running DDR1022 being an issue.

What mem voltage are you using? How much NBv/MCHv have you given? You might just have enough to run now but not enough to be stable...parts do degrade over time, some faster than others unfortunately as you know...so maybe a small bump up in voltages in this area would help.

Keep us posted on your results!

Right now its running at 2v as well for mine, and yea the 2x2gb blue kit. Its running a straight 1000mhz right now for the testing. The north bridge is at 1.4v right now and there's no reason why it should need anymore than that. I did try putting the ram to 2.1v but it didn't make a difference.


As for when i'm booting up and when it crashes, it seems to be when nothing is active at all strangely enough. It'll finish loading everything and it will just be the system idle process when things go down hill. Initially it made me think that a virus is a possible option as well although avg was unable to find anything (can run avg while everything is locked up and theres no issues). I'm probably going to look at reformatting in the near future regardless of what memtest comes up with, because something may have become corrupt.

[magibeg]

Its completed one full pass and there's no errors yet for memtest , I think i'll leave it going for a few more hours to be on the safe side though, next step would be a reformat i suppose.

[Kursah]

What are you running for a CPU OC on settings/voltage? 3.665/1.43v is currently correct? What FSB/Multi are you using to get there?

With OC'd Quads it does help to feed a little more NB/MCHv, FSBv, etc. I've had that kind of issue from a number of things, generally not a completely stable OC was the culprit, and the most probable route I've found for resolution was CPU OC. I personally run with EIST/C1E/Speedstep enabled, sure it can bring up issues, but I also like the thought of not being at full speed unless I need it. But when I had my Q I did go through similar issues, I finally settled on 400FSB, and a few different settings to finally get stability, I had RAM issues, heat issues, vdroop issues, the usual laundry list of overclocking that we do out best to qualm for stability.

I'd say run 400x9, 1.42v if the Q is stable at that voltage, DDR1000, 5-5-5-15, 2.0v, 52-54TRFC (iirc), 1.4vNB, 1.4vFSB. I guess it depends on what settings you're currently using, but I think it's that you're pushing a little too far unfortunately. There's instability beyond the memory that is causing your issues here I believe...and it could be a number of things...but if we can get 400FSB to run good and stable...we can work ya back up to your slightly higher OC.

I wouldn't quite reformat and re-install windows quite yet...keep playing with BIOS, and just step down to a better known OC level for Q6600's, see what we can get for stability out of it and go from there.

[magibeg]

Originally Posted by Kursah

What are you running for a CPU OC on settings/voltage? 3.665/1.43v is currently correct? What FSB/Multi are you using to get there?

With OC'd Quads it does help to feed a little more NB/MCHv, FSBv, etc. I've had that kind of issue from a number of things, generally not a completely stable OC was the culprit, and the most probable route I've found for resolution was CPU OC. I personally run with EIST/C1E/Speedstep enabled, sure it can bring up issues, but I also like the thought of not being at full speed unless I need it. But when I had my Q I did go through similar issues, I finally settled on 400FSB, and a few different settings to finally get stability, I had RAM issues, heat issues, vdroop issues, the usual laundry list of overclocking that we do out best to qualm for stability.

I'd say run 400x9, 1.42v if the Q is stable at that voltage, DDR1000, 5-5-5-15, 2.0v, 52-54TRFC (iirc), 1.4vNB, 1.4vFSB. I guess it depends on what settings you're currently using, but I think it's that you're pushing a little too far unfortunately. There's instability beyond the memory that is causing your issues here I believe...and it could be a number of things...but if we can get 400FSB to run good and stable...we can work ya back up to your slightly higher OC.

I wouldn't quite reformat and re-install windows quite yet...keep playing with BIOS, and just step down to a better known OC level for Q6600's, see what we can get for stability out of it and go from there.

Thanks, but the problem is that as a basic part of trying to get things stable again i went back to stock settings for everything and i'm still getting the exact same issues. Its really making me question if its the hardware or not when it doesn't seem to matter what settings i'm currently running i still get the exact same issues with instability. Its not even a random instability issue, its a very predictable one that happens every time at roughly the exact same time.

I've actually had my q6600 up as high as 3.8ghz with 1.45 volts with no trouble in the past but i didn't want to cut back on the life span. Unless i've buggered things so much that they don't even run properly at stock settings, but even still that doesn't explain why it would error out on the exact same thing every time without getting errors on virtually any test i'm able to run on it that checks for errors.

[magibeg]

AVG randomly found a virus and 'healed' it. So far no crashing in windows when back at my 'regular' settings. Will probably reformat anyway just to be on the safe side, don't like the idea of a possible virus chilling on my computer.

[sneekypeet]

Glad to see you seem to have sorted this out!

[Kursah]

That's interesting...I don't blame ya for wanting to reformat man, I try to keep that task a minimal situation anymore, but sometimes it's just gotta be done.

With the supposed virus gone have you gone back to your overclocked settings and given it a go?

[magibeg]

Yup, I'm running happily at 3.6ghz again just the way i should be. It was definitely a strange time, my first instinct was hardware instability, but i couldn't make it fail regardless of the torture tests i put it through on any of the hardware. The fact that it also seemed to be primarily windows affected (only froze the desktop/taskbar/internet explorer) seemed to point to some sort of software instability.

I do have one bit of advice though, if you're computer is used by other family members/kept in a common area, it may be wise to make separate accounts, like what my sister will be getting shortly as shes the one who downloaded the virus through limewire. Blackeyed Pea's did not release a song thats 134kb in size regardless of quality.

[Kursah]

ROFL! Yep I've been there with my G/F before lol...glad you got it up to par man!