Spot checking for faults - at my wits end

Servo said:
Try prime95 blend test. i've got the same setup and had problems like you.

Remove 4 gig for testing.

Up the volts on the core to 1.35, ram 2.1, nb 1.36. fsb term 1.6/7
Click to expand...

I'm running the default blend test now, if all is well i'll open my box later and remove some ram and retest. If nothing then i'll mess with the voltages.

Edit. Are those suggested voltages for testing or for normal running? I haven't overclocked my RAM if that make a difference.
 
Last edited:
Swi1ch said:
I'm running the default blend test now, if all is well i'll open my box later and remove some ram and retest. If nothing then i'll mess with the voltages.

Edit. Are those suggested voltages for testing or for normal running? I haven't overclocked my RAM if that make a difference.
Click to expand...

Try furmark before you bother ripping the RAM out. The reason I suggested linpack is because it causes the CPU to produce more heat than Prime95.

OCCT = http://www.ocbase.com/perestroika_en/ once you have it running the linpack stress-test option will be obvious. It will monitor temps & voltages for you (so you can observe to see whether any of the rails are dropping untowardly).
 
Your XMS2 should be set at 2.1v as standard.

I know this as I have 4Gb of the stuff and it does not like anything below 2.1v

Its a safe voltage :)
 
dustofnations said:
Try furmark before you bother ripping the RAM out. The reason I suggested linpack is because it causes the CPU to produce more heat than Prime95.

OCCT = http://www.ocbase.com/perestroika_en/ once you have it running the linpack stress-test option will be obvious. It will monitor temps & voltages for you (so you can observe to see whether any of the rails are dropping untowardly).
Click to expand...

Any tips for optimising the Furmark stability test? Do I want "Xtreme Burning Mode", "Log GPU Temp" or "Post Processing"?

Edit. Ran furmark for a while and no errors or anything, but it got me thinking; I'm not even getting in game, i'm crashing at the menu before anything intensive has rendered anyway.
 
Last edited:
Swi1ch said:
Any tips for optimising the Furmark stability test? Do I want "Xtreme Burning Mode", "Log GPU Temp" or "Post Processing"?

Edit. Ran furmark for a while and no errors or anything, but it got me thinking; I'm not even getting in game, i'm crashing at the menu before anything intensive has rendered anyway.
Click to expand...

So maybe your problems are elsewhere, or Driver based. Give OCCT linpack a try, then the other suggestions provided (memtest, driver cleaning utility + reinstall)

When OCCT outputs graphs (after it has finished running) perhaps you could upload them somewhere and post them here so that we can have a look
 
Last edited:
dustofnations said:
So maybe your problems are elsewhere, or Driver based. Give OCCT linpack a try, then the other suggestions provided (memtest, driver cleaning utility + reinstall)

When OCCT outputs graphs (after it has finished running) perhaps you could upload them somewhere and post them here so that we can have a look
Click to expand...

I tried ccleaner and reinstall and that didn't help, although I will try rolling back drivers as someone suggested.

Currently running a full blend test (didn't let it run the full two hours before) and if that comes clean i've got OCCT installed and ready to go.

I am wondering though, since I changed the recover setting to not auto-restart and go to BSOD instead, yet it still restarts, is this indicative of a fault?
 
Swi1ch said:
I tried ccleaner and reinstall and that didn't help, although I will try rolling back drivers as someone suggested.

Currently running a full blend test (didn't let it run the full two hours before) and if that comes clean i've got OCCT installed and ready to go.

I am wondering though, since I changed the recover setting to not auto-restart and go to BSOD instead, yet it still restarts, is this indicative of a fault?
Click to expand...

I don't think CCleaner is what you're after. Try Driver Sweeper?
 
ANDARIAL said:
As has been said before try upping the ram volts
What volts are you running the ram at?now
Click to expand...

I'm looking at the DRAM Volt setting; it's currently set to auto, but also says the default setting is 1.80. Is this the setting I should bump up to 2.10?

Oh, two hours Prime95 blend test no errors.
 
dustofnations said:
Yep bump it up, then try the things I mentioned earlier with OCCT and posting the graphs... Good luck.
Click to expand...

I'm running OCCT CPU Linpack test now. Under the monitoring tab, it provides 3 Voltage meters; VCore, 3.3V and 12V, is there anything I should be watching with those?
 
Swi1ch said:
I'm running OCCT CPU Linpack test now. Under the monitoring tab, it provides 3 Voltage meters; VCore, 3.3V and 12V, is there anything I should be watching with those?
Click to expand...

When it finishes it will output some JPEG files which will show the voltages over time amongst other things (5 or 6 images I think)
 
dustofnations said:
Can't be any harm. Do the problems still persist after the RAM voltage has been upped?
Click to expand...

I'm getting the same problems with the ram voltage pushed up to 2.10. I'll upload the images if I can find where OCCT stored them.

Isn't there another way to find out what is causing the crashing? Surely it'll be easier to address then.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom