safe voltage for a Q6600 (not G0)?

Harbz · 26 Jul 2007 at 11:40

I am having a problem with my overclock where the pc reboots itself after 30 mins or so when priming. Its a Q6600 running at 3.4G @ 1.42V underload. Temps are around 68C avg across the cores watercooled.

Bit worried to go higher on the volts. would a low vcore be causing the reboots?

The board is a gigabyte DS4 P965 V1.0.

Tooks · 26 Jul 2007 at 12:00

Could be a number of things, but volts is a common culprit.

I'm always happy to go to about 1.5v, but uncomfortable any higher!

I don't think the voltage increase would increase your temps though, and some of my mates think nothing of running their CPUs 24/7 at up to 1.6v....

It could be you've reached the limit of your CPU and no amount of voltage will make it stable at 3.4ghz.

ArchAnGeL · 26 Jul 2007 at 12:12

voltage increase = temp increase

Tooks · 26 Jul 2007 at 12:44

Agreed, but not as much as everybody thinks...

What I was trying to get across was that going from 1.42 to, say, 1.475 won't dramatically increase temps compared to where they are already.

I've only ever seen 2 or 3 degrees difference between 1.35 and 1.5 at the same FSB speed on my PC's.

Cob · 26 Jul 2007 at 16:49

Tooks said:
Agreed, but not as much as everybody thinks...

What I was trying to get across was that going from 1.42 to, say, 1.475 won't dramatically increase temps compared to where they are already.

I've only ever seen 2 or 3 degrees difference between 1.35 and 1.5 at the same FSB speed on my PC's.

The Quads are a different kettle of fish. Upping the vcore by 0.05v can have a severe effect on load temps.

Mine needs 1.42v under load (1.5v in BIOS) to be stable at 3.33ghz. Definitely give the vcore a nudge if you want 3.4ghz, but try to keep the load temps below 70c. And considering you're already at 68c, I would suggest you drop down to ~3.3ghz and be happy there.

Tooks · 26 Jul 2007 at 17:23

Cob said:
The Quads are a different kettle of fish. Upping the vcore by 0.05v can have a severe effect on load temps.

Are they that hot?!

I know that they're essentially 2 x core2's sandwiched in the same chip, but assuming you could take away all the excess heat effectively, why would the CPU internally be any hotter than one core2?

2 x a heat source at the same temp is the same as one x heat source isn't it? It would just take more cooling to take away the excess heat?

I'm not saying you're wrong by the way, just keen to understand why you're saying that quads don't behave the same as a standard core 2 regarding increases in voltage.

What was the temp of your CPU at stock FSB/stock vcore?

juno_first · 26 Jul 2007 at 23:10

I'd deffo up the voltage as 68c is nothing for a quad

Find the sweet spot just before the chip tells you to go to h e l l , mine is 1.525v, then overclock until prime gives an error or it blows up

mofnet · 27 Jul 2007 at 12:55

i have noticed that the quads seem to behave differently towards heat than duals.. my thinking is that one core is being heated by the other core sitting next to it.. if u think about it, its like placing a little heater along side the cpu core... its obvious i would think that this would create more heat. and its a recursive thing, the first core will heat the second and vise versa...

i use my rigs mainly for 3d rendering and i have noticed that over a period of say 2 to 3 hours the temps seem to "creep" very slowly up an extra degree or two before settling out at a steady reading, whereas my previous dual cores and before that single cores seemed to achieve their "maximum" temps within say 10 mins or so and then not move upwards at all.. so this creeping upwards for an hour or two i put down to having the 2 individual cores on the same die heating each other up....

i would think that this will change a bit once we're dealing with 4 cores inside of one die... perhaps?

cheers
ian

Tooks · 27 Jul 2007 at 14:20

I don't see how quads can be changing the laws of physics here though?!!

If somebody more knowledgeable than myself could explain things, then I'd be grateful!!

Sure there is a greater quantity of heat being produced, because there are twice the amount of cores than before, but assuming your chosen cooling solution is up to it, they shouldn't in theory run appreciably hotter than a standard core2?

I've noticed that temps creep up a degree or two on my pc's when they've been working all day, so I think that's normal and not just a quad thing. Other things in the case soak up heat and then radiate it and of course generally the room gets warmer too as the pc's churn out hot air.

To be honest, I think the quad hot running thing is being reported a bit hysterically at the moment, similar to the good old 2900XT...

I'm picking one up myself tomorrow, so I'll find out for myself!!

Yes, they run a bit hotter, but it's not by much, and anyway they've been designed to cope with it.

But then we knew that already didn't we?!

Chronictank · 27 Jul 2007 at 14:31

1.5 is as high as i would go personally, but you should be good as high as you can keep everything sufficiently cooled. I have not heard of a machine breaking down because of too high voltage provided proper cooling is in place you probably reduce the lifetime of the components but lets face it you would have upgraded long before that happens

Cob · 27 Jul 2007 at 16:48

Tooks said:
I don't see how quads can be changing the laws of physics here though?!!

If somebody more knowledgeable than myself could explain things, then I'd be grateful!!

Sure there is a greater quantity of heat being produced, because there are twice the amount of cores than before, but assuming your chosen cooling solution is up to it, they shouldn't in theory run appreciably hotter than a standard core2?

I've noticed that temps creep up a degree or two on my pc's when they've been working all day, so I think that's normal and not just a quad thing. Other things in the case soak up heat and then radiate it and of course generally the room gets warmer too as the pc's churn out hot air.

To be honest, I think the quad hot running thing is being reported a bit hysterically at the moment, similar to the good old 2900XT...

I'm picking one up myself tomorrow, so I'll find out for myself!!

Yes, they run a bit hotter, but it's not by much, and anyway they've been designed to cope with it.

But then we knew that already didn't we?!

I've yet to see an E6600 with decent cooling hitting 90c under load at 'only' 3.5ghz and 1.46v in Windows.

But, of course, I'm being hysterical, aren't I :rolleyes:

Tooks · 27 Jul 2007 at 17:26

Alright, no need to be sarcastic!! I'm just trying to understand what's going on here. Play nicely please!

The point I'm trying to get across is that there is twice the heat source with a quad, but those heat sources should be the same max temp as a standard core 2? There is just more volume of heat to shift, not necessarily a higher temp of heat to shift?

So, the problem would appear to be with the inadequacy of current cooling solutions which have been designed at most around dual core cpu's? What could be happening is that some coolers are being overwhelmed by the four cores, and are not able to dissipate the heat effectively or quickly enough, so temps climb leading to folks saying they're 'too hot'.

Remove the cooler from a core2 cpu, and then see what temp it gets too, I bet you'll easily see 90c before it shuts down. Both dual and quad core cpu's will get just as hot as each other, it's just that one is easier to cool than the other.

My view is that quads don't run that hot, it's just that not all current cooling solutions can cool them adequately...

I'll see when I put mine under water tomorrow.

EDIT:

Cob said:
I've yet to see an E6600 with decent cooling hitting 90c under load at 'only' 3.5ghz and 1.46v in Windows.

I presume that means you know of a quad hitting 90c under load, if so, I would respectfully suggest that the problem there is the cooling, not the chip running 'hot'! If I ran my core2 at 3.4Ghz under a passive heatsink, and got load temps of 90c, would all core2's be 'hot' then? You'd rightly tell me that it wasn't the chip that was 'hot', but that my cooling was inadequate...

There are two more cores here, so your cooling needs to reflect that. Saying they run hot is just too simplistic, it's like saying my car runs hot with no water in the radiator. It just wouldn't be being cooled properly...

Cob · 27 Jul 2007 at 18:13

You're doubting the experiences of people that actually own the chips. Sarcasm should be the least you can expect

My cooling's fine for my Quad (D-Tek Fusion, 120.2, North Atlantic gales). The warmer 2 cores sit at ~70c at 3.33ghz and 1.42v in Windows under load. I won't push it any higher because it's Folding under full load 24/7. My E6600 averaged about 64c under load with the same cooling at 3.6ghz and 1.52v in Windows.

It's easy to get a Quad to 90c. Enough owners are openly admitting on these forums that they're hitting 85-90c under stress-testing with decent cooling once they get to 3.5ghz and what would be considered reasonable vcore for the older C2D's. 50c at idle and stock settings with high-end air is also being reported by some owners. Hence why many people are waiting for the G0's.

cokecan · 27 Jul 2007 at 18:41

my Q regularly gets to 75 with an SI-128 and 140mm fan at stock on full load (folding also)

when im not folding it does idle at around 50 give or take a few degrees across the cores

you may be right in that the current cooling solutions arnt 100% up to the task of cooling the quads but then again, i dont see my temps as too hot?? sure their hotter than a duo but theres more cores. i mean its stable, doesnt give me any problems and theres still plenty of head room before it will want to throttle or shutdown, job's a gudden as far as im concerned

Chronictank · 27 Jul 2007 at 20:09

Tooks said:
So, the problem would appear to be with the inadequacy of current cooling solutions which have been designed at most around dual core cpu's? What could be happening is that some coolers are being overwhelmed by the four cores, and are not able to dissipate the heat effectively or quickly enough, so temps climb leading to folks saying they're 'too hot'.

No, coolers are designed around cooling a metal plate, the number of cores in the design doesnt really have a whole lot of bearing on anything tbh
Your point may hold about them not cooling the processor sufficiently if it is a poor cooler (such as using a c2d stock cooler on a quad), but lets face it you wouldnt be overclocking with a sub-standard cooler in the first place so the point is mute.

The only exception possibly being watercooling where the design of the waterblock may influence its cooling power slightly, but even then the difference would be negligble as you are performing the same process, worst case one/two of the cores would be running hotter than the others and that would be very noticable and would indicate a problem with the cooling method.

Tooks · 28 Jul 2007 at 10:02

Cob said:
You're doubting the experiences of people that actually own the chips. Sarcasm should be the least you can expect

My cooling's fine for my Quad (D-Tek Fusion, 120.2, North Atlantic gales). The warmer 2 cores sit at ~70c at 3.33ghz and 1.42v in Windows under load. I won't push it any higher because it's Folding under full load 24/7. My E6600 averaged about 64c under load with the same cooling at 3.6ghz and 1.52v in Windows.

It's easy to get a Quad to 90c. Enough owners are openly admitting on these forums that they're hitting 85-90c under stress-testing with decent cooling once they get to 3.5ghz and what would be considered reasonable vcore for the older C2D's. 50c at idle and stock settings with high-end air is also being reported by some owners. Hence why many people are waiting for the G0's.

My apologies if I came across as doubting the owners of such chips, that wasn't really what I was driving at!

I'm just trying to understand the reasons why some are getting these high temps.

Without doing any rigorous study, it would appear from the couple of replies in this thread and elsewhere that people on water don't seem to be getting the same very high temps, lending a bit of weight to the 'some air coolers being overwhelmed' theory? Afterall, water 'scales' much better and has the capacity, usually, to shift more heat more efficiently?

Nebzor · 28 Jul 2007 at 11:28

Tooks said:
My apologies if I came across as doubting the owners of such chips, that wasn't really what I was driving at!

I'm just trying to understand the reasons why some are getting these high temps.

Without doing any rigorous study, it would appear from the couple of replies in this thread and elsewhere that people on water don't seem to be getting the same very high temps, lending a bit of weight to the 'some air coolers being overwhelmed' theory? Afterall, water 'scales' much better and has the capacity, usually, to shift more heat more efficiently?

A is dishing out 150 units of heat energy, B is dishing out 150 units of heat energy. You're expecting A and B to be dishing out 150 units together, when you say they should be as hot as C2Ds.

Tooks · 28 Jul 2007 at 13:35

No, I'm not saying that at all.

To adapt your example...

If A (core2duo) is dishing out 150 units of heat energy, and your cooler can shift those units, then you get a given load temp...

If B (Core2quad) is dishing out 300 units of heat energy (as it's effectively 2 x core2's lumped together), and your cooler can shift those 300 units, then you should have a given load temp similar to a core2duo...

That's only IF your cooling solutiion can shift 300 units of heat energy though.

If it can't, then you'll get higher temps, same as any undercooled cpu.

That's all I'm saying, that in theory, and given the correct cooling solution, your temps should be not much higher than a core2duo, as evidenced by those on water (which is generally a higher capacity cooling system) seeing far less issues with temps.

Core2quads create more heat, not necessarily a higher temp heat, so they are not simply 'hot'.

Nebzor · 28 Jul 2007 at 13:45

Tooks said:
No, I'm not saying that at all.

To adapt your example...

If A (core2duo) is dishing out 150 units of heat energy, and your cooler can shift those units, then you get a given load temp...

If B (Core2quad) is dishing out 300 units of heat energy (as it's effectively 2 x core2's lumped together), and your cooler can shift those 300 units, then you should have a given load temp similar to a core2duo...

That's only IF your cooling solutiion can shift 300 units of heat energy though.

If it can't, then you'll get higher temps, same as any undercooled cpu.

That's all I'm saying, that in theory, and given the correct cooling solution, your temps should be not much higher than a core2duo, as evidenced by those on water (which is generally a higher capacity cooling system) seeing far less issues with temps.

Core2quads create more heat, not necessarily a higher temp heat, so they are not simply 'hot'.

Well, they are hot, because you're not going to find an adequate air cooler (adequate in the fact it can fully remove the heat the second pair of cores is producing). You're just being pedantic.

Tooks · 28 Jul 2007 at 14:08

Well, if you can't see the point I'm trying to make, then I'm sorry.

Pedantry is not my style, I was just trying to have an adult discussion. :rolleyes:

Coolers have always had to adapt to new generation CPU's, and it's no different for quads.

I was just challenging the assumption that quads are 'hot', because before you know it, everybody on here will be taking that as gospel, which I don't believe to be true. All they need is the right cooler, same as any cpu.

Actually, for the speeds they run at and considering how many cores they have, quads actually run pretty cool, especially in comparison to previous generation netburst chips.