Degration, 65nm chips, on what volts?

[Drinkyoghurt]

Hey guys,

I have a e6420, 65nm of course and I was wondering how far I could push the envelope with the voltage

I used to run this cpu @ 3.6ghz 1.55v a while back, I was wondering how far I could cut back on the voltage and got it down to 1.47v

that's great and all, but now I run 3.7ghz @ 1.52v and I was wondering if running my cpu at those volts all the time would degrade it a lot(like let's say, almost depleted in 6 months) or will it still last a good year or two?

i've heard about 45nm e8400's degrade within a week of running over 1.4v and that's what made me wonder

 

[Spawn-Inc]

not having alot of experience (only had 2 overclocked cpus b4) my cutoff for 24/7 is 1.45-1.5v. i would guess at 1.47 you could get a year or more, but thats a complete guess.

 

[Vargtass]

Been feeding my X6800 1.60 for over one year by now - no degration.
It's a B3 - think its need a bit VCore.

*first post *

 

[muddocktor]

I also think that temps play a big factor in the equation too. The cooler the chip runs, the longer it will take the abuse without degradation. That's why I try to keep my temps as low as possible on my farm with premium hsf's.

 

[Drinkyoghurt]

the load temp is max 54c

my processor craps out @ 60c with a nice BSOD, so I try to keep it under 58c

the only time it got 60c was when I didn't turn on the fan for my TRUE

 

[Jayws]

I personally prefer to stay below 1.4vCore and 55c load with 60c being the absolutely acceptable max. If it's some older tech though that has lost value and needs a kick in the butt to get me to my next upgrade, well... my standards loosen up a bit.

 

[goog`]

well i had my e6400 at 1.68 and it degraded in a few months now I have a celly at 1.7 and i'm sure it will soon follow so >1.67 is bad for sure

 

[Sam__]

Been feeding my X6800 1.60 for over one year by now - no degration.
It's a B3 - think its need a bit VCore.

*first post *

For 24/7 voltage I wouldnt go above 1.55 and 65-70C.

 

[hokiealumnus]

Thought I'd throw my $.02 in here. My E4400 is the first chip I've overclocked and it's been folding @ 3.2GHz (320 x 10), 1.445Vcore (1.41v after Vdroop) for seven months without a hiccup. It's taken some abuse too...benched for several hours at 3.5GHz, 1.56v (actual) and then for Forum Wars, for about an hour'ish, it benched at 3.6GHz, 1.60v. Reset the OC back to 3.2GHz/1.41v and it's been folding along ever since.

 

[w2richwood]

i agree with not going over 1.4 and 60c temps to be on the safe side of things less you have the funds to replaces thing quiet offen
Rich

 

[maxfly]

ive been running my 3070(e6700)at 1.475v 33c/47c for over a year.zero degradation problems for me to date.

 

[satandole666]

My E6600 never saw below 1.5v since I got it about a year and a half ago. Its still humming along fine at stock now in my HTPC, but if anything my overclocks got better with time. I started out with a cake 3.6ghz and worked it up to 3.8ghz on water.

I wouldn't worry about anything sub 1.6v tbh, but that's just me.

 

[Jayws]

Just curious... When you have a performance related issue, how do you know that you're looking at "degredation"...

 

[Albuquerque]

I think the old-school rule-of-thumb has been +15% as the effective limit.

So if your stock VID is 1.2375 then ~1.423 is about all the further you should consider going. Your mileage may vary

 

[satandole666]

I think the old-school rule-of-thumb has been +15% as the effective limit.

So if your stock VID is 1.2375 then ~1.423 is about all the further you should consider going. Your mileage may vary

Wasn't the max VID for the E6600 1.375v? If so then ~1.581 would be 15% and that sounds reasonable to me.

Intel specced (wish I could find the document) 1.45v as the max for 45nm before they claimed lifespan would be drastically effected. No way 65nm would be less than that.

Either way, 1.58 is still a bit high. If your budget can't afford to replace a dead chip don't risk it and be happy with what you can get with 1.4-1.5v.

 

[MadMan007]

I like to stick to my totally non-scientific and unfounded rule of thumb of no greater than 10% above rated max vcore. So for a 1.35V Max CPU I'd go up to 1.485V for example. No real rhyme or reason to it other than a 10% variation may be engineered in, probably it's more like 5% but we ARE overclocking here

 

[Albuquerque]

Wasn't the max VID for the E6600 1.375v? If so then ~1.581 would be 15% and that sounds reasonable to me.

Intel specced (wish I could find the document) 1.45v as the max for 45nm before they claimed lifespan would be drastically effected. No way 65nm would be less than that.

Either way, 1.58 is still a bit high. If your budget can't afford to replace a dead chip don't risk it and be happy with what you can get with 1.4-1.5v.

My original values were there only for "example" purposes. I used 1.2375 because that's the VID of my Q9450

 

[Automata]

Just curious... When you have a performance related issue, how do you know that you're looking at "degredation"...

When you have to up the voltage to stay at the same speed. I had that happen with my E6400 when I had it. It ran 1.55v @ 3.6ghz. Eventually it was 1.65v to run 3.6ghz. At that point, I just backed it way off and started looking at something else

 

[Big Mike]

So far my rule of thumb has always been decreasing returns and temps, my E6600 ran for about 9 months at 1.52ish @ 3.7 ghz, it would creep to almost 3.8 but took almost 1.6v to get there and was in the mid 60s fully loaded, I only went that high to bench and never had any issues with it. I figure whatever lifespan shortening I may have incurred will ultimately be offset by the fact that the CPU will be completely worthless before it won't run at its stock speed and voltages. It runs great at 1.35 and 3ghz right now in the rig im building it into to sell.

 

[duper]

I don't think I will be touching my volts anytime soon.