Crashing problems (q6600 + 8800GT)

Hi guys.

I have a Q6600, an asus P5K vanilla, and a 512mb 8800GT

I've had the Q6600 at 3.4ghz (425x8) since about November. I was getting intermittent crashing in games (never in windows), but it wasn't too frequent so I didn't bother to resolve the issue.

Incidentally, I ran prime95 on four cores for hours, and it was largely OK. One core did actually mess up eventually, after quite a few hours, and I think I upped the vcore slightly (core voltage is 1.432v idle, as I write this - it's high because my mainboard has large vdroop, and under load the core voltage drops quite a lot).

So, I thought the CPU was fairly stable. It ran for hours and hours under 100% on each core.

Anyway, lately the crashing seemed to get worse. I had two crashes in the space of about 12 hours, so I lowered my FSB to 400x8 giving me 3.2ghz. It seemed fine, until the PC crashed again about 5 minutes into a game of Supreme Commander.

Now I'm more confused. CPU or Graphics problem?

The crashes are usually just a lockup, the screen just freezes on the current thing it's rendering (although the last one is the exception to this, the screen went black, which has never happened before).

Obviously I should try to set my CPU and mobo to stock, and test it. But in the meantime, can anyone tell me what this seems most like? Do CPU instability problems cause lockups like this in games only?

I'm using 8800GT's latest drivers. Graphics card doesn't seem to be running too hot (probably under 80 degrees).

My q6600 didn't seem to like the 8 multi, but you need to post all your bios setting. you didn't mention what voltage you have set your north bridge to. your running 4 gigs of mem so you might need to up the NB volts. and maybe try a 9 multi to get your overclock. where is your memory set?? is it over clocked?? need more information to help you out.

Sounds more like a graphics problem. Do you have another card you could test?

I have no other graphics card to test with.

@gottcha, I've collected all of my BIOS information. I also set everything to the defaults for now, to see if this fixes anything.

So, here are the settings that I was using to run 3.4 (and 3.2ghz by changing the FSB to 400 from 425), which have been crashing in games intermittently:

GeIL is my RAM manufacturer, and other than the FSB overclock, they're running at stock timings of 4-4-4-12.

I actually have the same exact problem with my Q6600.

It was flat out stable at 2.88ghz but, I wanted more, upped the OC to 3.2ghz w/vcore of 1.32 and an fsb of 1.4

Ever since that OC, I've been beyond unstable.

Ram checks out to be perfect but, I just can't seem to find the problem.

Every blue screen I get says Hard Drive cannot be read and blue-screens me, my guess is I'm getting corruption because of crap timings so, tomorrow I'm ordering a new set of reapers...will see how that fairs

See, I don't even get any blue screens or anything. I just get a crash, and my event logs say "unexpected shutdown". I get no information.

If it was CPU or motherboard, would I be more likely to get BSODs? Or do my game freezings seem like graphics card problems?

I'm currently running back at stock speeds. I'm gonna run for a week or so, and see if I get any more crashes in games. Still, it wouldn't necessarily prove it, as it's sometimes gone for a period as long as a week with no crashing anyway.

So any suggestions for my BIOS settings, assuming that my problems are down to my overclock?

from your bios settings it seems you have your cpu voltage set at 1.5?!?!?!!?!??!?!? that really high for 3.4ghz i think i was running 1.3280 in windows to reach that speed. if your air cooling whats your temp at?!?!??! use core temp the newest one to get the temp. the cpu voltage referance i didn't need to reach 3.4 maybe put that in auto(when i did try to change this it brought my temps up about 5-7 degrees). i don't have your board so these are just recommendations. and maybe try moving north bridge volatage to say 1.4 volts that what i had to have to get 3.4ghz
keep memory at stock speeds untill you conferm your chip is stable
here are my settings for 3.4ghz

multi=9
fsb=386
vcore bios 1.34375
vcore under load 1.3280
cpu pll 1.50 (default)
fsb term voltage 1.38
dram voltage 1.96 (mem takes 2.2volt but p5e overvolts)
nb volts 1.4
sb volt 1.075

if your don't see the setting it in auto

try it out and let me know.
hope it helps.

like i said i had been useing the 8 multi and it would give me random reboots and bsod's, but every set up is different

What kind of powersupply are you using, and can you use a volt meter to measure voltages when at load? It is difficult to diagnose a problem with crashing if you never really did any extensive stability testing when initially setting your overclock.

corruption said:

What kind of powersupply are you using, and can you use a volt meter to measure voltages when at load? It is difficult to diagnose a problem with crashing if you never really did any extensive stability testing when initially setting your overclock.

Click to expand...

he should be able to use "cpu z" to check what his voltage is on load.

gottcha said:

he should be able to use "cpu z" to check what his voltage is on load.

Click to expand...

I take all software readings with a grain of salt....I was also referring to 12v, 5v, and 3v rails on his powersupply.

corruption said:

I take all software readings with a grain of salt....I was also referring to 12v, 5v, and 3v rails on his powersupply.

Click to expand...

NO need to bite my head off!!!! Just kidding!!! i thought you were talking about his vcore under load. i agree he needs to list all his part to get a good answer.

@corruption: My PSU is a 580W Hiper jobbie. I figured it should be OK was it was quite expensive and the reviews seemed OK.

@gottcha: I know 1.5v in BIOS seems really high... but there are two reasons it's that high:
1) The offset between the BIOS setting and the CPU-Z reading is quite large - so at 1.5v BIOS the idle CPU-Z reading was 1.43v or so. (V-offset I believe this is called?)
2) 1.43v is high too, but under load it was dropping low, because of V-droop. At anything lower than 1.5v in BIOS, it was getting as low as 1.27v sometimes under heavy load.

Anyway, I thought I'd sorted all this voltage stuff... I figured that I'd set my Vcore high enough for 3.4ghz. I'm not convinced my problems are CPU voltage related - I was running for like... 5 months on 3.4ghz, and games were crashing now and again. But after dropping to 3.2ghz and keeping the same CPU voltage settings, I still crashed. If it was because it's not getting enough juice, then I'd have expected the 3.2ghz setting to go a bit longer without crashing up my PC.

I have no multimeter, and I've not performed a pencil mod on my motherboard (lots of people do a pencil mod to change the value of a resistor, which apparently helps with Vdroop).

If problems are CPU voltage related, I'm screwed anyway - I can't really push the voltage up safely as it's already at 1.5 and load temperatures are about 60-62 degrees. I'm not willing to do a pencil mod, I think I'd rather just replace the motherboard than do that.

Hopefully though, there's something else amiss with my bios settings. Like, maybe my NB and SB voltages.

gottcha, you wrote:
cpu pll 1.50 (default)
fsb term voltage 1.38
nb volts 1.4
sb volt 1.075

All of my settings for that are Auto. So I should try those to get my 3.4ghz back. However, do you know what the default values for those are, when set to auto? That'd help give me some perspective on what I'm doing.

Thanks.

I did some searching and found that this might help.

cpl pll - 1.5
fsb voltage - 1.3
NB volts- 1.4

try that and see if it helps. any setting in auto is not good for over clocking.
as for you not wanting to do a pencil mod............. let me just say its EASY, CHEAP, WORKS, it also has the power to get rid of headaches and frustration due to overclocking with the droop.


from seeing you volts go from 1.5 (bios) to 1.43 (windows) to 1.27 (under load)WOW thats bad. mine is set at 1.34... in bios and windows 1.3280 and load 1.3280(pencil mod)
if it drops to 1.27 under load that could be why your overclock is failing.

Ok, well I tried those settings, and got a BSOD about 10 minutes into running OCCT.

Here's a nice graph of my vcore, courtesy of OCCT:



I don't know if this is the problem, since I also crashed at 3.2ghz (400fsb).

Also, OCCT was fine back in november, at 3.4ghz settings. Maybe it's hotter now? Load temps are around the 63-68c region (which doesn't seem THAT bad).

The first thing I would suggest is to set everything back to stock speeds and see if the problems persist. I have experienced situations where Prime95 is 12+ hours stable but games will crash randomly. After returning everything to stock the crashing ceased, so you really can't rely on a single stability testing application to determine if your overclock is 100% stable.

start said:

Ok, well I tried those settings, and got a BSOD about 10 minutes into running OCCT.

Here's a nice graph of my vcore, courtesy of OCCT:



I don't know if this is the problem, since I also crashed at 3.2ghz (400fsb).

Also, OCCT was fine back in november, at 3.4ghz settings. Maybe it's hotter now? Load temps are around the 63-68c region (which doesn't seem THAT bad).

Click to expand...

I would have to say its your vcore causeing the crashes it seems to be dipping to low under load.

http://forums.anandtech.com/messageview.aspx?catid=28&threadid=2059762&enterthread=y

try the percil mod!!!!

Yeah, I guess it is probably just the voltage. Damn.

KillrBuckeye, yeah I tried the stock speeds and everything seemed OK. I was going to run it for a week, but I'm impatient, and it worked fine for a couple of days. It's looking more and more like it's the overclock.

I think I'll try running 425mhz x 7 for the next few days. I'll try that at the Vcore settings I was using for 3.4ghz, and presumably the vdroop will be a lot lower at 3ghz than at 3.4ghz (even though dropping to 3.2ghz didn't stop the problem). Obviously 425mhz is still quite high, but it'll be a good test of the mobo and RAM.

If that proves stable, then I guess I should have a bash at the pencil mod (OR upgrade to a better motherboard?) ... But after I try 425x7. That should give a good indication of whether it's a voltage problem I hope.

It seems to me that the next logical thing to do would be to use the 9x multi to take the FSB and memory speed out of the equation. Otherwise you'll still have no idea if it's insufficient Vcore, VDimm, or chipset voltage causing your problem. If everything is okay at 9x355 (3.2 GHz), then at least you know it's not the CPU causing your instability.

KillrBuckeye said:

It seems to me that the next logical thing to do would be to use the 9x multi to take the FSB and memory speed out of the equation. Otherwise you'll still have no idea if it's insufficient Vcore, VDimm, or chipset voltage causing your problem. If everything is okay at 9x355 (3.2 GHz), then at least you know it's not the CPU causing your instability.

Click to expand...

+1

and do the pencil mod!!!!! i was very ify about doing it but once i did i would do it to every board i own. its great. it might also take a couple of time to get it right but its well worth the 10-15 min it will take.