Help Needed: Orthos + e4500 + Vista 64 -> Random restarts on blend

melbourne720 · 27 Feb 2008 at 10:16

Not sure what you mean mate.

Something is stopping this chip from clocking stabley. I doubt it is the upper limit of the motherboard, but I would be happy to eliminate that (hence the 290 x 7). At the moment my guesses are (from most likely to least):-

CPU Voltage (Vdroop, drop, or not enough period)
FSB Voltage
Northbridge Voltage
Strap problems
Northbridge cooling
CPU cooling

Clairvoyant · 27 Feb 2008 at 10:37

Ok I may need to go and play.

What would you say is a safe Vcore (droop under load is aout 0.2)

Temps are low to mid 50s at the mo.

melbourne720 · 27 Feb 2008 at 11:46

I wouldn't go above 1.5 without water cooling. When BIOS set VCore is 1.4, I am getting a little Vdrop (CPUZ reads 1.39) but quite a bit of Vdroop (1.37-1.36 under load) so I may need to up it 1.45 in the BIOS to get at or above 1.4 constant.

melbourne720 · 27 Feb 2008 at 16:57

FSB is stable:

Next step: I have increase the vcore to 1.45, the VVT from 1.3 to 1.4 and am going for 300 x 10x , Small ffts 8hrs +. Then try for 3.2

melbourne720 · 28 Feb 2008 at 08:05

300 x 10 was fine, 14hrs small FFTs stable. Will try for 3.1 or 3.2 next. Think the extra VCore (possible VVT) might have done it.

TechnoCorruption · 28 Feb 2008 at 10:55

Your doing well with the results now, if it was me though I would try again with the blend test at 300x10, as far as I know small ffts really only stresses the CPU and the cache inside it.

Blend should be the test to see if your initial problem has been solved, then once that passes then I would start upping the cpu again. The only problem you have with the way you are doing it is that if you get a fail now your not going to know if its the cpu holding back or the same problem you had initially.

Just my opinion of course

melbourne720 · 28 Feb 2008 at 22:44

Yeah, Blend at 300 x 10 sounds like a plan.

Thanks for all your help Agr3sive, remind me to nominate you for MoH at the end of the year

melbourne720 · 29 Feb 2008 at 07:22

TechnoCorruption · 29 Feb 2008 at 10:07

That's a great result there Melbourne, at least now you have this as a fall back should any further overclocks be unsuccessful, it does look like it was the low set strap that was giving you the errors before.

You should be good on the 266 strap for the rest of the overclocking on the CPU, mine is doing 380fsb on the 266 strap at present. How much more are you going to try on the CPU?

My E2160 needs 1.55v bios (1.52V Windows-Load) for 3.42GHz but some may not be willing to go that high on the volts for 24/7, in my opinion its safe but I wouldn't use more than 1.55v on a 65nm for 24/7.

And your welcome for the help mate, I know how annoying it is to come back to the computer during a prime/orthos session and see that it has failed

melbourne720 · 29 Feb 2008 at 14:00

I'm letting the above run for 24 hrs, as it is a build for my brother and I want a 24hr stable speed. Once I have that in the bank, I'll see what more I can get out of it at sensible temps. 3.2 would be nice.