Max safe temp for E7400

[MetalX]

I ran some stress tests (Prime95) on my newly overclocked E7400 at 4074 MHz and 1.48v, and after about 9 hours of testing the CPU was perfectly stable, but the core temperature as listed by coretemp was 76C. That seems rather hot for a CPU core. So that leads me to ask: What is the highest safe temperature for an E7400?

 

[kimsland]

Maximum case temperature at the maximum Thermal Design Power (TDP) value: 74.1°C
From here: http://processorfinder.intel.com/details.aspx?sSpec=SLB9Y#

 

[MetalX]

Oh ok, that's a lot higher than I thought. So I guess 76C at 4.07 GHz isn't so bad then

 

[kimsland]

Try this:

Reseating your CPU

 

[MetalX]

I know, I do need some new thermal paste, but 76C was with my CPU fan on low. Putting it to high would lower the temps a good bit.

 

[kimsland]

And make it noisy

 

[MetalX]

True, but it's all I can do until I can get my hands on some quality thermal grease. I'm using the stuff that came with my CNPS 9500 right now.

I ran out of silver grease.

 

[kimsland]

I've always wondered how much of a real difference it would make just using any ol' grease. I suppose no one has really tested this fully

 

[MetalX]

I read an article a few years back that found that the classic AS5 that everyone swears by is actually about equal to everything else on the market. The article found that it doesn't seem to matter what kind of grease you use, as long as there's something between the CPU and the heatsink.

I wish I could find the article I think it was from MaximumPC...

 

[captaincranky]

You Two are Sort of a Disappointment.........

I will probably live to regret replying to this the way I'm going to but, fools rush in......

It's probably my failing, but this thread makes no sense to me whatsoever. MetalX, you have more than 1800 posts and you really should know better. You're trying to manipulate the answers to align with you want to hear.

First kimsland tells you this;

Maximum case temperature at the maximum Thermal Design Power (TDP) value: 74.1°C
From here: http://processorfinder.intel.com/details.aspx?sSpec=SLB9Y#

Then, you answer with this;

Oh ok, that's a lot higher than I thought. So I guess 76C at 4.07 GHz isn't so bad then

I'm certain both of you can read and count with the best of them. Still, 76C is higher ( and therefore hotter) than 74.1C. So, the CPU is running hotter that it's design spec. Now, take a moment to explain why that's, "not so bad".

We've gone through this silliness about stock vs. AS5 (and other) after market thermal compounds many times. And i believe that we've come to the conclusion that save for couple degrees, they pretty much all work the same. In other words, if you buy the best thermal compound you can find, there isn't going to be a miracle in the offing that drops your CPU temp to even 65C. All other things being equal.

Now, an E7400, fine CPU be though it may, is not an I-7. No amount of overclocking and overvolting is going to turn it in to one. You'll just break it. If that's not important, let 'er rip.

You've already got one of the better HSFs on the market. So, IMHO, turn the clock down, and drop the Vcore. You won't have as much to brag about, but the machine will probably live longer.

Plan "B" is water cooling. It always seemed stupid to me to go through maybe $300.00 worth of air coolers, only to find out that they all overheat, then think about water cooling.

Has everything else been optimized with respect to case cooling?

I'm thinking that operating at about 70 to 80% of maximum CPU temp would be considered within design parameters. (In the mid sixties) Besides, when the CPU gets that hot it should be throttling back. That would place the available CPU power back at the levels it was before it was overclocked anyway.

 

[MetalX]

I will probably live to regret replying to this the way I'm going to but, fools rush in......

It's probably my failing, but this thread makes no sense to me whatsoever. MetalX, you have more than 1800 posts and you really should know better. You're trying to manipulate the answers to align with you want to hear.

I'm certain both of you can read and count with the best of them. Still, 76C is higher ( and therefore hotter) than 74.1C. So, the CPU is running hotter that it's design spec. Now, take a moment to explain why that's, "not so bad".

Ok, I shall

74.1C is the maximum temperature specified by Intel. 1.3625v is the maximum voltage specified by Intel. A desktop CPU is designed to last for a fairly long time, I'd guess at least 5 years or so of near constant use. So if Intel designs the CPU to last 5 years, I would think that any maximum temperatures or voltages they specify would be the maximum the CPU could run at for that entire 5 years without shortening its lifespan (theoretically).

I do not expect my CPUs to last 5 years. I upgrade every 12-18 months. If my CPU only has to last 12-18 months, why not maximize the performance for those 12-18 months?

I maximize the performance by overclocking it. As you know, overvolting reduces the life of the CPU, and increases the heat. Heat output scales geometrically with voltage. I realize that 1.48v is a good bit over the "maximum" of 1.3625v. Now I have no proof whatsoever to back this up, but I made an assumption that my CPU would last at least a year or two running at this voltage assuming it was not overheating.

But running at such a high voltage (and to a lesser extent, from a thermal perspective, clock speed) causes my CPU to reach 76 degrees after approximately one hour of constant 100% load. I would assume that running the CPU at a high temperature consistently would reduce its lifespan, in addition to the lifespan loss of overvolting, so I wondered first what the manufacturer specs had to say about the highest temperature the CPU should run at. But I assume that operating within the manufacturer specs will cause my CPU to last 5+ years, therefore I believe that I could probably run it a little hotter than that, so long as it remains stable.

So I decided to ask Techspot for their opinion on what temperature would be "safe". Not safe as in Intel specifications, but "safe" from the perspective of someone who only needs the CPU to last a year. I probably should have specified that in my original post.

Kimsland responded with the pure Intel specs, which I admit, I had not searched for. The Intel specs were not really the answer I was looking for, but they allowed me to use my experience to get the answer I was looking for (namely, whether 76C will kill my CPU in a year or not). I decided that the temps I'm getting are safe.

We then degenerated into a conversation about thermal grease that probably should have been done in IM or something like that.

Now, an E7400, fine CPU be though it may, is not an I-7. No amount of overclocking and overvolting is going to turn it in to one. You'll just break it. If that's not important, let 'er rip.

Keeping my CPU "healthy" is not of utmost importance to me. Performance and bang for my buck is. You could argue that if I blow the CPU, the bang for my buck is pointless because I now have a broken CPU. I recognize the validity of this arguement, but I simply do not care, and I will not care until (if) I do blow it. I'm immature like that.

You've already got one of the better HSFs on the market. So, IMHO, turn the clock down, and drop the Vcore. You won't have as much to brag about, but the machine will probably live longer.

I think I've covered that.

Plan "B" is water cooling. It always seemed stupid to me to go through maybe $300.00 worth of air coolers, only to find out that they all overheat, then think about water cooling.

I could use watercooling, and I may do so in future.

Has everything else been optimized with respect to case cooling?

Almost. My case (Antec Nine Hundred) has one more empty 120mm fan bay, which I may populate one of these days.

Does this clarify my reasoning for you?

I understand my "argument" is full of holes, personal opinions and assumptions, but I'm trying to explain to you my reasoning

 

[captaincranky]

This is Purely Hypothetical......

OK. I'm on board with all of this but, I am at a loss to figure out upon which facts one would use as a basis to extrapolate how much overvoltage would be allowable to shorten the CPU life to a year. This is silly but, if that could be done, arguably one could decide to use a little less Vcore ,so that the CPU would burn up in 2 years. As you pointed out in another thread (loose quote, about CPU overclocking), "individual results may vary".

In any event, I did say "let 'er rip". I meant "let 'er rip". Still, as I understand it, when Intel CPUs hit 70C @ 100% usage they throttle back to 70% usage on their own. This, (if I'm correct) does negate the 30% overclock and you're back where you started.

The only problem I'm having is one of "duty cycle". Very few things are designed to operate at 100% duty cycle. Well OK, top fuel dragsters are. But the motors only last maybe 6 miles. Meaning, couldn't you see it in your way to run the poor E7400 @ about 3.6Ghz and 65C. It would love you for it.

I am an old fuddy-duddy though. so my E7300 (stock speed)in an Antec 900 runs @ about 35C, fans on low. Boring, I konw.