Errors on Prime, but stable :confused:!

rumble291

rumble291

rumble291

Associate
Joined
4 Mar 2010
Posts
139
Hey guys,

I got my new gear about a week ago, I got a new CPU cooler (Noctua NH-D14) its a beast and keeps my CPU at like 50c Max load with a good 25c ambient temp! Currently stress testing at 45c has been for a few hours now.

The main problem is that I keep getting errors on a single core but then all cores run fully maxed out for hours and hours max i've tested is 6 hours and one core failed.

Could this be due to voltages? Too high or too low?

Atm im running 3625Mhz at 1.4875 Vcore. I know the Vcore is pretty high :( I've been testing it and it seems to be running fine except for the error on prime 95. I've played games, left my computer running for a full day without any issues :confused:.

Is there anything else that could be affecting this? Bare in mind that i've upped my FSB to 250Mhz so that everything such as RAM, HT link and NB stays at an even speed and can increase if needed. I've used the multiplyer to get to 3625Mhz.

Thanks guys,

Kurt
 
could be your nb, ht link speed, whatare they running at - you may need to lower the nb & ht multiplier or increase the nb mvolts other than that it could be memory.
 
If an Error, then it's problem with OC settings!

Bare in mind,.. if one game haven't failed it doesn't mean that other game won't ... game can stop responding after couple hours of play!
 
Is it the message...FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress .txt file.

I had that on mine and after hunting around on forums i found out it was the mem timings, lower the spd on the mem, it may sort it out!!
 
Its more likely to be down the the HT link being overclocked by ~25% as this is the link between the IMC and the CPU dropping the HT Link and North bridge back to stock frequency should solve your problems.

Tight timings wont be an issue as the memory should be running at stock mhz
 
That error is almost always to do with the RAM! I'd slacken the timings and increase a bit of voltage. If you still can't get it working properly, run it through memtest for a few hours and see if anything crops up. (at stock values, that way you know if the ram is defect)
 
rounding error is usually ram, or ram controller. solution to the former has been mentioned, solution to the latter is to up the cpu-nb volts a bit.
 
Ok thanks guys, still being ****ty with me! went for 1 hour 45 mins before a fatal hardware error :S other cores still fine mind you.

CPU-NB is 1.2500 volts that alright or add a bit more? HT Link voltage is 1.26v
added a bit to the NB aswell 1.14 all running at 2500Mhz
 
ive got my nb/ht multiplier at x9 and 2355mhz. (my board is sposed to support 2600) nb @1.2375 mv on my asrock mobo. ive also stuck a zalman heatsink onit and have a large fan pushing a lot of air passed it. as it was getting quite hot - around 60-65c now around 50c

i got that error a lot when setting up my oc. it can be memory or ht/nb speed. increasing the mv will help.

you need to drop your speed a little 2500mhz is on the limit. you might need also to check your memory isnt the problem at that speed to rule it out at least.
 
Last edited:
upped the CPU/NB voltage to 1.2625 and it did almost 7 hours straight, before an error occured on OCCT hopefully its just another notch up on the voltage and should do me for the solid 9 hours.

Cheers for the help guys,

Kurt
 
You must log in or register to reply here.
Back
Top Bottom