Q6600 crashing all the time but prime95 no errors?

Hey guys,


Got myself a Q6600 and really.. my system has been crap ever since...

After installing i installed new 7.8 Ati drivers for vid card and went away to play supreme commander... after about 40 mins, the screen went blank, but game was still running... after a reboot everything was 4bit 640x480.

Un-installed the display drivers and then went to 7.6 catalyst (was originally running 7.2 or something) and even removed ATI Tray Tools, and stopped overclocking my card...

The display went back to normal but now when im playing a game (GTR2 Supreme Comm) it will crash and come up with "[email protected] has encountered an error and needs to close" or at worst the PC has even simply rebooted and get a windows message saying "Windows has recovered from a serious error".

I figured that this may be the CPU either not being seated properly, or maybe getting too hot... but I have been running Prime95 (running 4 concurrent prime95 actually) for over an hour and not a single error, and also the temps are a little high, but still acceptable (70 70 67 67 in coretemp.. rock solid) which makes me believe that whatever the error is, its definitely not the CPU or else it would have shat itself 6 times over by now.

So... im thinking that the problem must be the X1950GT... what should I do? how could I tell, and why would it all of a sudden just spaz out?

Im going to try even older drivers to see if thats the problem, but is there a way i can test if it is a hardware problem without swapping the cards? (i dont know anyone else with an AGP card lol)

Anyway, i guess the graphics card would be a good excuse to upgrade the motherboard and RAM to something with PCIe and DDR2... but Im glad to know that its not the Q6600 thats killing things.

ATI native drivers are crap IMO, install the Omega ATI drivers, much better

What are your GPU temperatures? Get ATItools and check.

What is your cooling solution and what FSB are you using?

i thought it was the other way around - that the core temps were higher than the IHS surface temps?

In any case, if it was cooking the chip, surely the thermal throttling would be cutting down the FSB??? but its still running happily at 2.4Ghz (2393.93MHz 265.99x9 to be exact)

MikeyD360 said:

i thought it was the other way around - that the core temps were higher than the IHS surface temps?

In any case, if it was cooking the chip, surely the thermal throttling would be cutting down the FSB??? but its still running happily at 2.4Ghz (2393.93MHz 265.99x9 to be exact)

Click to expand...

DUH, I'm asleep, yep its reversed.....I'll go to sleep again. 70c is still pretty hot, are you on stock cooling?

orion456 said:

DUH, I'm asleep, yep its reversed.....I'll go to sleep again. 70c is still pretty hot, are you on stock cooling?

Click to expand...

Yeah, stock cooling - you had me worried for a moment there - and stock HSF wax too, lol.. too lazy to put my AS on...

Also my case fans are all on medium.

I certainly wouldn't be overclocking the CPU with these temps, but at least I know its not the CPU that crapping out.


stupid stupid vid card... i'll try messing around with drivers, maybe even a new XP install and see what happens.

BTW, Prime95 still no errors...

Also - temps have actually gotten lower over time.. settled into 68 69 67 66 - i guess i havent really run the CPU 100% for this amount of time since i installed it, so probably still burning in the wax.


Oh, the other thing... when I do video encoding (ie DVD to Xvid) it will also error out of that after an hour or so... how can it be so crap for everything else, but run sweet in prime95?

First of all, is your system overclocked? Prime95 only test the CPU stablilities, Not the RAM. The reason why i am asking this because the only time your computer restarts when you play games for a while; that concludes your system is unstable (northbridge, southbridge might be too hot too). Why dont you try this stress program OCCT, this program stress CPU and RAM (i prefer this over Prime95). Run the 30 mins test twice, if you pass both. Then it prove 99% your system is stable. Oh ya, post your bios setting if you cant get it fix.

It could also be the psu that cant keep up, if its mem problem prime will find it.

well the PSU is brand new 600w, so should be able to handle stock clocked system.

ok, well i tried that OCCT program and crashed after like 3 mins... so what is the problem?

Prime95 will run CPU torture test for over an hour with no difficulties, OCCT crashes after 2 minutes and real world software crashes anywhere between 40min and 2 hours when intensive tasks...

Where do I look?

I'm betting your system is too hot.

1. take the side panel off and get a house fan, blow some air in there, do the temperatures drop significantly? If yes, you need better air flow into that case.

2. What is the CPU voltage? What ever it is, turn it up a bit, and try again. Keep the house fan blowing.

3. How hot is your case with the side panel on?

Its not the case.. haha... its an Antec900... if airflow is a problem in this case then theres no hope for any of us...

You're talking about a case with 5x 120mm fans, and 1 200mm fan...
theres a 200mm fan poking out of the top tight above the CPU, theres a 120mm fan at the back of the case next to the CPU, theres a 120mm fan in the middle of the case blowing over the northbridge and VGA card, 2 120mm fans at the front of the case blowing over the HDD and into the case and lastly on the side is a 120mm fan sucking air from the VGA area out.

the air coming from all the fan areas is cold as ambient and the fans suck/blow in a uniform direction to keep airflow moving from front to back over the important areas.

OCCT ran for 13 mins that time and then crashed... for some reason, Prime95 heats up the CPU to 69C but OCCT only gets the temp up to 57C so i have trouble imagining CPU temp as the problem if it crashed in both scenarios.

Ok, with all that air, you should be as cool as possible, though the noise must be annoying.

How about underclocking to FSB 200 and turn the voltage down to 1.25v or even less if you can. Does it still hiccup?

Hmmm, is that a sound card? Maybe remove that and see what happens.

Your video card design blows all that hot air down on the mobo? I like the ones blowing out the back of the case better. An x1950 creats a lot of heat, I wonder if it is causing some overheating on the mobo itself?

According to your mobo specs, "FSB1066-CPU is supported by overclocking only when you adopt an external VGA card and DDR400 CL2.5 memory modules". Is it possible your memory isn't up to the task? Do you have CL2.5 memory?

http://www.asrock.com/mb/overview.asp?Model=775I65G

Why did you pick such a small board to go in a big case?

Try running orthos with the ram stress option

It was the only board I could find that had AGP8X support for my X1950 and for DDR400... looking back now it may have been better to hold off and upgrade whole thing, but the video card is tough to replace... by the way, It is very good of you to actually look up stuff on the web to try to assist me, I appreciate it.

If i am correct.. i think one of my DIMMs is CL2.5 and the other may be a 3... would that make such a thing to happen?

I am leaning towards memory now because when i do the Prime95 CPU only, it never faults.. when I do CPU and some RAM, or BLEND then it faults...

*edit.. the ram module both say CAS Latency 2.5@166MHz and 3.0@200Mhz.. CPU-Z shows them running at 177.3Mhz.. so how does that fit into anything?

MikeyD360 said:

I am leaning towards memory now because when i do the Prime95 CPU only, it never faults.. when I do CPU and some RAM, or BLEND then it faults...

Click to expand...

Then that is a sure sign of memory failure

Remove the cas3 stick and see if it passes.

MikeyD360 said:

It was the only board I could find that had AGP8X support for my X1950 and for DDR400... looking back now it may have been better to hold off and upgrade whole thing, but the video card is tough to replace... by the way, It is very good of you to actually look up stuff on the web to try to assist me, I appreciate it.

If i am correct.. i think one of my DIMMs is CL2.5 and the other may be a 3... would that make such a thing to happen?

I am leaning towards memory now because when i do the Prime95 CPU only, it never faults.. when I do CPU and some RAM, or BLEND then it faults...

*edit.. the ram module both say CAS Latency 2.5@166MHz and 3.0@200Mhz.. CPU-Z shows them running at 177.3Mhz.. so how does that fit into anything?

Click to expand...

Ah, those compromises can be a killer.....


They have to be at 166mhz to run at 2.5, so you are out of spec and its causing random errors. Time for a board and/or memory upgrade I think or else set the FSB lower to get the 166mhz spec. What kind of memory is it, specifically? Or maybe set the FSB:RAM divider to get the same result.

The ram is pretty poor.. its like Hyundai or something...
The MoBo is also pretty poor... hence, no way to manually change the ratio.. only option is adjustment to the FSB, everything else changes along with that...

MikeyD360 said:

The ram is pretty poor.. its like Hyundai or something...
The MoBo is also pretty poor... hence, no way to manually change the ratio.. only option is adjustment to the FSB, everything else changes along with that...

Click to expand...

Yeh, not really a Quad friendly board and that ddr400 CL2.5 memory seems expensive from the big boys.

So the FSB goes down then until it becomes stable. Hopefully the voltage goes down as well and your heat will become better.

Time to save up for the new board and memory to make that quad fly!