Associate
- Joined
- 27 Feb 2006
- Posts
- 1,750
- Location
- Normanton
as above just don't want to push too much through it trying to reach 5Ghz
Max safe voltage for a i3 under water
- Thread starter mattyj1085
- Start date
I have had a look but can't find much info...
From just looking at rjkoneill o/c results he started to get diminishing returns vcore - mhz at around 1.4 - 1.45v I would imagine then given the way other past CPU have performed and the 45nm tollerances I would imaging this would be the 'safe-ish' place to stop..
From just looking at rjkoneill o/c results he started to get diminishing returns vcore - mhz at around 1.4 - 1.45v I would imagine then given the way other past CPU have performed and the 45nm tollerances I would imaging this would be the 'safe-ish' place to stop..
- Joined
- 27 Feb 2006
- Posts
- 1,750
- Location
- Normanton
Cheers I thought 1.4v ish would be safe I will see what I can get out of her tonight then
Honest answer is we have no idea. Not been out long enough for any to die. It's dropped to 32nm, it's going to be more vulnerable to voltage than 45nm was. 1.5V or even 1.6V was considered fair game at 65nm, at 45nm advice was to stay below 1.45V and preferably under 1.4V. From this you could guess 1.3V limit, but it would just be a guess.
Mean time before failure scales linearly with processing size, 65nm should last twice as long as 32nm at the same voltage. I don't know how current density varies with voltage, but you could take a good stab at estimating it by finding the power draw of the i3 chip. Someone has probably recorded power draw of a 45nm i5 chip and of a 32nm i3 chip in the same motherboard, with luck at a stated voltage. Unfortunately current density is probably a function of process size as well, which brings me back to "who knows". I'm going to check intel's maximum vid.
vid range is 0.65V-1.400V which suggests intel think it'll be just fine. This range is often taken as the "safe" range by people on the basis that intel wouldn't release a chip running at an unsafe voltage. The flaw there is that running 3ghz at 1.4V draws less current than 4ghz at 1.4V, so to stay true to "intel knows best" one would have to scale the maximum voltage down as frequency increases.
Mean time before failure scales linearly with processing size, 65nm should last twice as long as 32nm at the same voltage. I don't know how current density varies with voltage, but you could take a good stab at estimating it by finding the power draw of the i3 chip. Someone has probably recorded power draw of a 45nm i5 chip and of a 32nm i3 chip in the same motherboard, with luck at a stated voltage. Unfortunately current density is probably a function of process size as well, which brings me back to "who knows". I'm going to check intel's maximum vid.
vid range is 0.65V-1.400V which suggests intel think it'll be just fine. This range is often taken as the "safe" range by people on the basis that intel wouldn't release a chip running at an unsafe voltage. The flaw there is that running 3ghz at 1.4V draws less current than 4ghz at 1.4V, so to stay true to "intel knows best" one would have to scale the maximum voltage down as frequency increases.
Last edited: