Would you run your Northwood at 1.7V+ ?

[cnhoff]

I have had some stability problems with my 1.8A at 1.7V, so i went higher, to 1.75V cooling setup, i get about 43C under full load, that is not much, i guess. So would you advise against running at this voltage?

 

[PliotronX]

Personally I would keep it below 1.7 and 1.7 maximum if I get a significant boost in overclocking. I would not risk electromigration just for a few 3DMarks.

 

[Mikki]

Agree. And depending on what board you have, your actual voltage will not match what you set in the bios. For instance, Asus boards generally overvolt, Epox boards generally undervolt. Make sure you're checking your voltage in Windows, not in your bios...

HTH 🙂

 

[cnhoff]

Seems like my 1.8A@2.4 runs more stable with 1.725 instead of 1.7. Maybe i will keep this setting, i have good cooling, running Prime for a few hours now at 44C.

 

[Mikki]

If you're just testing that's cool, but if you are gonna keep it around that speed I would highly suggest you back off the fsb rather than up the v-core that much....😉

 

[gururu]

yes, drop the vcore as the understanding is that higher voltages expedite the decay of the copper conduits.
it is not a matter of one day your chip works, one day it doesn't. The chip will get slower with time, even at a
constant clock speed.

 

[AndyHui]

It's definitely bad to run over 1.7 volts.

 

[Barrei]

At 1.725v your fine ,once something is stated on a forum alot of people think it's gospel , even at 1.75v the chances of electromigration are minimal .

 

[Duvie]

I agree with barrei...I have seen some say 1.65 to 1.75v...It is all personal preference. I stay right near 1.7v....


I haven't to date seen anyone loose one for staying 1.75v or less. It is when ppl go into 1.8v territory and then deal with board specifics as Mikki pointed out that can lead to some damaged chips...I can speak form experience on this and the chip does not just die. It slowly degrades to the point of ot excepting as much voltage and thus severly limiting the oc. I was able to downclock mine at the hint of trouble and found new stability at 1.8@2.25 w/ 1.58v...It ran there for 10 weeks until I got this 1.6a....


What mobo? Those vcores sound suspiciously like quoted bios settings.

 

[malandro]

My experience has been that after a .1v increase in Vcore you reach a point of diminishing returns. Stay within a +.1v increase and you are still within design specs so you should be ok. If its not going to work at 1.6v chances are that adding more volts will not help that much. If I were you I would go down to 1.6v and see what happens. I know it sounds anathema but give it a try and you might be pleasantly surprised. I have tried 4 different 1.6A cpu's and everyone of them went to 2.3Ghz without a hitch at 1.6v.

 

[Duvie]

My experience has been that after a .1v increase in Vcore you reach a point of diminishing returns. Stay within a +.1v increase and you are still within design specs so you should be ok. If its not going to work at 1.6v chances are that adding more volts will not help that much. If I were you I would go down to 1.6v and see what happens. I know it sounds anathema but give it a try and you might be pleasantly surprised. I have tried 4 different 1.6A cpu's and everyone of them went to 2.3Ghz without a hitch at 1.6v.

Wrong!!!! The whole above statement is riddled with inaccuracies....

Go read INtels own pdf on the northwood!!!

If its not going to work at 1.6v chances are that adding more volts will not help that much.

LOL!!! Hmmm...I guess then when I was at like just under 2.6ghz with 1.60v I should have stopped cause I couldn't get much better..HUh??? Read sig!!!

My experience has been that after a .1v increase in Vcore you reach a point of diminishing returns.

What evidence do you have of this??? Just cause benchmarks do not scale linearly doesn't mean this is true...My 2.736ghz at 1.71v seems to be scaling quite properly...It actually increases in scaling in snadra and pcmark2002 benchmarks due to the fact the system fsb keeps raising with each boost in fsb....

I have tried 4 different 1.6A cpu's and everyone of them went to 2.3Ghz without a hitch at 1.6v.

This doesn't mean sh^t....What were the packing dates on all of these processors? What were the mobos they were ran on?...All factors!! blanket statements check at the door....



Edit the above post...you don't sound like you're qualified to speak much!!!!

 

[gururu]

there is no 'chance of electromigration'. the chance is 100%. it occurs anytime a current is run through a circuit. metal atoms will always be displaced in the current, resulting in corrosion. heat and voltage will indeed expedite the process of decay. the end result is more electrical resistance along the conduit; affecting the speed of the operation. it is impossible to say at what voltage or temperature and under what time frame a noticeable loss or defect in performance will occur.

my only advice: either play conservative (1.5, 1.6) or go for the gold (1.7 and above).
As for myself, I went for the gold. but I usually advise people to stay under 1.7 if possible.

 

[Acanthus]

i read a thread here somewhere that said over 1.85v and supercooling can kill a P4 in under 2 weex