core i5 running at 101 degrees C, as if nothing...

Started by
24 comments, last by JohnnyCode 9 years, 6 months ago

Hey gamedev,

I'm running Cubemapgen (modified by Sebastien Lagarde), and I notice if I change the "input degamma" slider it runs some seriously heavy algorithm that runs for like 20 minutes (what the heck ??).

Anyway, during this time it brings my little core i5 at 98C for core0 and 101C for core1.

its a 3230M supposed to run at 2.6 Ghz stock, but coretemp reports its actually at 3Ghz. It has a T-Junction of 105 degrees so the core circuitry itself is safe, but you know, at 100 it may melt stuff around !

Its a Dell 17 inches laptop.

What do you think ?

Advertisement

I think you should do something about it. Those temps are too hot, so underclock it back to its recommended specs or something, or check the thermal paste (if at all possible on a laptop) and the fans. If all else fails, maybe try a cooling pad. Laptops have a very narrow operating range due to their cooling limitations, and running at 100+C is not recommended, not to mention uncomfortable as it'll heat up everything around it, including the keyboard.

I would say from experience that most laptops run 45 to 60 degrees C idle, and 70-85 on load. But 100C is just too hot!

(that said I have not owned a laptop for several years - maybe they are more tolerant to heat, or are just meant to run hotter these days)

“If I understand the standard right it is legal and safe to do this but the resulting value could be anything.”

Yes, I use it as a desktop actually, it is on the side, mounted on top of wood cubes and it has a 120mm fan blowing under it, for the hard drive originally (because it was clicking).

On idle, intel speed step is throttling it at 1.2 Ghz, 0.83Volts and it is at 60 deg.

The fan seems off for all I can tell, because the hard drive noise is stronger than the fan.

Correction: it just started spinning as I write this. (but to no change of temp)

I have noticed that, as an Ivy bridge, starting cubemapgen by default was using the internal intel 4000 graphics card, so the chip was really at its fullest utilization.

Apparently when I use the nvidia card instead it goes up to 85C, the fan accelerates slightly.

Its true that the keyborad becomes hot as hell, but I use a desktop keyboard on the side, so its not like I'm constantly aware of the laptop's temp. Which is scary lol.

The weird part is that it doesn't clock down, when the turbo thing hits in, it reaches up to 3.2 Ghz, then its normally at 3.0. But I'm sure this CPU is just a 2.6 CPU so what the hell ?

The power plan mentions "100%" under "cpu power", not 115%. meh anyway..

i5-3230M

What you are seeing is Turbo Boost kicking in.


Intel® Turbo Boost Technology dynamically increases the processor's frequency as needed by taking advantage of thermal and power headroom to give you a burst of speed when you need it, and increased energy efficiency when you don’t.

And since you haven't crossed its "danger zone" yet, it is trying to squeeze as much speed as it can.

Since you are running a very CPU intensive algorithm, the processor is ramping itself up to give you a little more oomph. My Lenovo laptop (also an i5) has several power option that can control performance and fan speed while on battery and plugged in. You might see if your Dell has something similar.

"I can't believe I'm defending logic to a turing machine." - Kent Woolworth [Other Space]

Yes definitely turbo boost thingie. But I would not expect it to wait for 105 degrees before spinning back to factory frequency.

I found a trick, I set the "maximum power state" to 99% in the power plan, and it runs a 2.6, at 85/89 degrees.

Less dangerous for surrounding parts.

How so "for surrounding parts"?

Printed circuit boards are usually rated for at least 180C. Remember they need to handle hot solder and handle thousands of heating/cooling cycles without damage. It is quite rare for companies to go less than that, and when they do the results are usually clear: The X360 Red Ring of Death was caused by a low temp rating on their PCB as they were bidding on the cheapest board they thought they could get away with.

I suppose if you are on an aluminum PowerBook it might be a problem as the metal conducts it toward more sensitive parts, but I think you would have mentioned it. Laptop manufacturers learned lessons from that incident, too.

Probably the most fragile of the "surrounding parts" is your own flesh touching the box. 100C is bad for skin, but the electronics can handle it.


Yes definitely turbo boost thingie. But I would not expect it to wait for 105 degrees before spinning back to factory frequency.

The T-Junction is just the maximum temperature allowed at the processor die.

See Intel Turbo Boost, it explains some of the conditions that need to trigger to enable and disable it.

"I can't believe I'm defending logic to a turing machine." - Kent Woolworth [Other Space]

The clock speed is definitely turbo boost, but you definitely should not be seeing temps of 100 degrees, and the processor should really only get that hot if you have insufficient or faulty cooling. Be sure the air intake and exhaust ports aren't blocked, check to make sure the fan and heat-sink are free of dust and debris, and blow them out with compressed air. You may have to partially disassemble the laptop to get at the CPU heat-sink (usually in a laptop, the heat-sink consists of a heat-pipe to some small fins near the exhaust port -- while you're in there, make sure that it looks like the heat-sink is seated properly. Also try running the laptop in a normal orientation rather than on its side -- the laptop is designed to work in a certain orientation; a good design ought not be so affected, but its possible the orientation is contributing to the problem and it'd be good to at least eliminate that as a possibility.

Its possible that this particular workload might put the CPU into some kind of pathological state -- either the CPU design in general, or perhaps your particular CPU due to a manufacturing defect. If the above doesn't help, your only recourse is to accept it and keep a close eye on it, try to get a warranty replacement, or perhaps you could use a CPU tweaking tool or the bios to under-clock/under-volt or be more aggressive about throttling itself a bit sooner (though, these options will give up some performance).

By the way, a normal load temp with sufficient cooling should be between 80-90 degrees C, although it depends somewhat on the ambient temperature.

Good luck, and let us know if any of these suggestions help, or if they've made no difference.

throw table_exception("(? ???)? ? ???");


but you definitely should not be seeing temps of 100 degrees, and the processor should really only get that hot if you have insufficient or faulty cooling.

Even, as he described, under heavy load continuously for 20 minutes?

Note that only one of the cores is getting warm, and that one core is less than the max chip temperature. My i7 hits 100C under heavy load occasionally, then throttles back.

I don't see it on CoreTemp, but from their web site it looks like it only measures DTS. While the raw values from DTS are nice, the PECI measurements show when the CPU's thermal controls kick in.

My guess is that the processor is running right up to the thermal limits and the PECI delta would be zero, because really, why not? As long as the cooling system is working you might as well run at full speed.


Even, as he described, under heavy load continuously for 20 minutes?

It really shouldn't from everything I know -- Heck, most overclockers recommend to keep temps under 90 degrees -- I mean, keep in mind we're talking centigrade here, 100C is literally hot enough to boil water. Its probably not a horror to hit it occasionally and then throttle, but any sustained time at that level would be enough to concern me that something's wrong with my cooling.

When my liquid cooling loop failed earlier this year, the CPU temp would peg itself at 100C by the time I made it to the windows desktop, the CPU was throttling itself all the way down to 800Mhz, and it would then crawl a bit higher before forcibly shut itself off after a few minutes -- a few degrees above 100C is actually the thermal tripping point for modern Intel CPUs, if it were to go much higher than OP is saying he sees and stay there, its got to be very close to turning itself off.

throw table_exception("(? ???)? ? ???");

This topic is closed to new replies.

Advertisement