Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/606
Article Thumbnail

Impact of Temperature on Intel CPU Performance

Written on October 28, 2014 by Matt Bach
Share:

Introduction

If you’re looking for authoritative information on how much cooling is enough for your CPU, you’ll be hard-pressed to find an answer. While we all know that modern processors need some degree of active cooling, very little official information exists to say how different temperatures affect a CPU's performance.

Older CPUs would simply fail if they started to overheat, but modern CPUs adjust their frequency based on temperature (among other things) to prevent a dramatic failure. Because of this, it stands to reason that once you reach a certain temperature, you will no longer be getting the maximum performance from your CPU because it will be busy protecting itself.

But what is that temperature? And do you really need a high-end liquid-cooled system to get peak performance, or is the little stock cooler that comes with most CPUs enough? In this article we will answer these questions and more.

How CPUs should be affected by temperature

Modern CPUs are able to adjust their operating frequencies through a number of technologies in order to either reduce their power consumption or provide maximum power as needed. The first of these modern technologies is Turbo Boost (or Turbo Core for AMD APU/CPUs). According to the overview page for Intel Turbo Boost Technology 2.0, there are five factors that affect the amount of increased frequency an Intel CPU can achieve via Turbo Boost:

  1. Type of workload
  2. Number of active cores
  3. Estimated current consumption
  4. Estimated power consumption
  5. Processor temperature

For this article, #5: Processor temperature is the main factor we are concerned about. You would expect Turbo Boost to slowly stepping back the amount of Turbo Boost as the CPU gets hotter. In actuality, an Intel CPU under heavy load will actually run at the maximum Turbo Boost allowed by the other four factors until it hits the CPU's thermal limit.

CPUs also have extremely robust thermal protection. If the CPU starts operating above the CPU's thermal limit it will begin to reduce the frequency in order to prevent catastrophic failure. Oddly, we have found that the thermal limit for both Turbo Boost and thermal protection on Intel CPUs to be right at 100 °C - which makes it very convenient to remember. In other words, until the CPU hits 100 °C you should see 100% of the CPUs available performance. Once you starting hitting 100 °C, however, the CPU will start throttling back to keep itself from overheating

How cooling actually affects Intel CPU performance

We know from both experience and explicit testing that modern Intel CPUs (at the very least Sandy Bridge, Ivy Bridge, and Haswell) can run at their maximum Turbo Boost frequency all the way up to 100 °C. While there may be a tiny performance difference between a CPU running at 30 °C and one running at 95 °C, our testing has found that the difference is miniscule. In fact, even after running benchmarks dozens of times the difference is so small that it is essentially nonexistent.

What we can measure is what happens once a CPU starts to hit 100 °C. To figure this out, we took an Intel Core i7 4790 and cooled it with a Gelid Silent Spirit Rev. 2 CPU cooler that was connected to a manual PWM fan speed controller. By running Linpack (which is a CPU benchmark widely used in the scientific community) and slowly dialing the fan speed down in careful increments, we were able to allow the CPU to overheat by incremental amounts. At each cooling increment we kept a log of the Linpack benchmark results as well as using CoreTemp to record the CPU core temperature and frequency.

Since the Intel CPU thermal limit is 100 °C, we can quantify the amount of overheating by measuring the amount of time the CPU temperature was running at > 99 °C.

CPU frequency based on cooling

We were very surprised when our testing showed that while the minimum CPU load frequency started to drop as soon as the CPU hit 100 °C, the average CPU frequency didn't drop by more than .1GHz until the CPU was overheating more than 30% of the time. In fact, Intel CPUs are surprisingly good at being able to handle this much heat with such a small reduction in the average frequency.

While this is a great way to understand what is happening from a frequency standpoint, we wanted to know how this affects real-world performance. To do so, we recorded the Linpack performance results to see what the actual performance implications of overheating are:

CPU performance based on cooling

Linpack is one of the most consistent CPU benchmarks we have ever seen and the Intel Core i7 4790 should score right at 190 GFlops with the problem size we used for this testing (30000). What is interesting is that the Linpack results from our testing almost exactly follow the average CPU frequency from the previous graph. While it makes complete sense, this pretty much just confirms that from a strictly CPU performance point of view, the performance of a CPU is directly related to it's average frequency.

So what does this all mean? Beyond the fact that Intel CPUs are impressively stable even while technically overheating it means that you can expect full performance from an Intel CPU as long as you keep it below 100 °C. At the same time, even if the CPU occasionally hits 100 °C you shouldn't see more than a minimal drop in performance until it spends a significant amount of time (more than 20% of the time) above 99 °C.

Is stock cooling adequate?

While the information from the previous section is really interesting, it may be hard to translate into a real world situation. To help with this, we performed the exact same Linpack test except that instead of altering the cooling manually we used two different CPU coolers - the stock cooler that came bundled with the Core i7 4790 and the budget-friendly Cooler Master Hyper TX3 which only costs about $20. To make this as real-world as possible, we installed our test hardware (Asus Sabertooth Z97 Mark II, Intel Core i7 4790, 4x Kingston HyperX LoVo DDR3-1600 4GB, NVIDIA GeForce GTX 980) into a Fractal Design Define R4 chassis with the stock fans running at 5V.

Intel stock cooling

As you can see from the graph above, even a very affordable CPU cooler like the CM Hyper TX3 is able to dramatically lower the CPU temperature under load. While the stock cooler is easily hitting 100 °C during the benchmark run, the CM Hyper TX3 only ever hits a maximum temperature of 80 °C. In terms of how much the CPU was overheating, the stock cooler allowed the CPU to run at > 99 °C about 20% of the time. 

Intel stock cooling performance

According to our graphs from the previous section, we should expect a measurable, albeit small, drop in performance with the stock cooler. We found that using the stock cooler made the CPU perform about 2.5% slower than it did with the CM Hyper TX3. The neat thing is that this drop in performance almost exactly lines up with our graphs of expected performance from the previous section.

Conclusion

The results of our testing can pretty much be summarized with the following three points:

  1. Modern Intel CPUs run at full speed (including the full Turbo Boost allowed based on the number of cores and workload) all the way up to 100 °C
  2. Even after the CPU hits 100 °C, the performance is not greatly affected until the CPU spends about 20% of the time > 99 °C
  3. While stock cooling only causes around a 2.5% drop in performance, even a budget after market cooler will dramatically improve CPU temperatures

Intel stock coolerFrankly, we were a bit surprised at how well modern Intel CPUs dealt with really high temperatures. They manage to run at full speed all the way up to 100 °C, and even then the performance is not greatly affected unless they spend a significant amount of time at that temperature. We certainly don't advocate letting your CPU run at those kinds of temperatures, however. While this article is about performance there are plenty of non-performance based reasond to keep your CPU temperature at a more reasonable level.

Sensitive electronics like CPUs have a finite lifespan and running them at higher temperatures shortens it. So unless you want to have an excuse to upgrade your system often, higher temperatures are counter-productive.

With PC hardware, higher temperatures make both minor and major hardware faults much more likely. These hardware faults can result in anything from reduced performance due to minor errors needing to be corrected to data corruption or bluescreens due to more dramatic errors.

The difficult thing is that just saying "lower temperatures are better" is an overly simplified way to look at CPU temperatures when you consider a PC as a whole. While you can use giant liquid cooling loops, insanely high flow fans, or even things like liquid nitrogen to keep a CPU extremely well cooled those methods are either expensive, cumbersome, loud, or a combination of all three. In reality it is generally better to let the CPU run a little warm in order to properly balance the cooling against the heat output of the CPU.

CM Hyper TX3

For example, the Corsair Hydro H60 is a great closed-loop CPU cooler that we use on most of our high wattage CPUs, but it would be overkill for a low-wattage Intel Core i3 CPU. It would certainly keep the CPU well cooled, but it would be much louder, more complex (due to the liquid and pump), and more expensive than what is necessary. Instead, a CM Hyper TX3 or Gelid Tranquillo (if you want to have the system even quieter) would be a better match for that CPU even though the CPU would run slightly hotter under load.

For the average system, our rule of thumb at Puget Systems is that the CPU should run around 80-85 °C when put under full load for an extended period of time. We have found that this gives the CPU plenty of thermal headroom, does not greatly impact the CPU's lifespan, and keeps the system rock stable without overdoing it on cooling. Lower temperatures are, of course, better (within reason) but if you want a target to aim for, 80-85 °C is what we generally recommend.

Tags: Intel, Turbo Boost, cooling

Hello Matt Bach

Thank you for your article.

I recall reading a number of years ago (20+) and perhaps not holding true today, that the average temperature (T) of a CPU (lithographic scale then being in the microns not nanometers) increases as the cube (^3) of the core clock frequency (f) such that [ f^3 = ~T ].

Thus, with linear increases in the core base frequency, the duty cycle (time) spent by a CPU above 100 C should increase non linearly as well.

In short, curious as to what core frequency (temp) your test CPU's, performance gain curve, rolls over when over clocked.

Park McGraw

Posted on 2014-10-31 01:03:56
LukeLetellier

I'd be interested to hear your thoughts on temperatures for long term XEON CPU.

I do a lot of 3D rendering - so it's normal for me to have my CPUs running at 100% for 50-75 hours straight to complete a project, and my initial thought is that having the temps at 80-85 degrees for that long (and using them this intensely for 3-5 years) would have negative consequences.

Posted on 2015-03-25 20:17:41

That is a really good question, and a hard one to answer. I'm going to make some assumptions and guesses here, so don't take this as anything more than just my opinion:

I think your CPUs will be completely fine. 80-85C sounds really hot, but remember that these CPUs are designed to run full speed until they hit 100C, so Intel sure seems to think that 80-85C is well within tolerance. I also looked up our failure rates for Xeon CPUs over the last four years and it is currently sitting right at .3% out of >2000 processors. That's really low and the latest generation of CPUs are even better than the old ones. In fact, in the last year we've only had a single Xeon CPU fail (for a .12% failure rate).

Honestly, I think you are more likely to run into motherboard problems long before you run into CPU problems. Supplying that much voltage constantly can be hard on the motherboard voltage regulators. Cooling the CPUs further will help a bit (as will making sure the VRMs on your motherboard are adequately cooled) but I think the motherboard is definitely going to be your point of most likely failure.

Posted on 2015-03-26 17:02:48

great article

Posted on 2015-04-17 17:19:24
nanook_northpole

My Intel i7/920@2.67 GHz desktop had thermal shut down yesterday after 5 years. Had to vacuum out all the dust & now it's speedy like it used to be !
Lesson: Vacuum out PC once a year !

Posted on 2015-04-21 06:33:36

One tip, based on what I have heard over the years: it may not be a good idea to vacuum *inside* a PC. I've read and been told that it can cause static electricity to build up, which could potentially damage electronic components. Instead, the prevailing wisdom seems to be that using air to blow dust out of the computer is best. Once out, it should be safe to vacuum (from the exterior of the computer, or if it falls on the floor, etc).

Posted on 2015-04-21 15:48:55
nanook_northpole

Once in circuit, electronic parts are rarely damaged by static [ I won't explain why ]. Don't worry about the forced air direction - just get rid of that dust ! And that story you heard sounds like marketing spiel from compressed aircan sellers !

Posted on 2015-04-21 16:59:38

Fascinating - I wonder why that advice has built up in the computer geek world then. I've seen and heard it all over, and a quick Google search uncovers it on advice boards like this:

http://www.howtogeek.com/57...

I wouldn't use a can of compressed air myself, though, except maybe for small, specific cleaning. I use an air compressor at home, and I think we have a large one here at work for cleaning out systems during repairs.

Posted on 2015-04-21 17:17:05
Mark Mywords

I agree with Nanook. Once in circuit, the static across any component will never build up.So vacuuming inside should cause no harm. Make shure you use the brush attachment, rather than a hard nozzle. That way you are less likely to physically damage a component. If the brush is carbon fibre, then there is even less likelihood of generating static.

When I vacuum, I also use compressed air. That is the only way of getting all of the dust, which is trapped in between the fins of the CPU heat sink. It also shifts the dust which the vacuum cleaner cannot reach.

Think of it as brushing and flossing.

Remember to also clean out the dust in the PSU. That will need to be blown out. I certainly wouldn't recommend opening that up.

After doing this to my PC after a year, it was then running so quietly, I could no longer hear it. This was summertime with an ambient temperature approaching 30C.

Posted on 2015-05-06 15:27:59
Germz

I use a vacuum. I agree that once in circuit, static is less likely to cause damage to hardware. Todays electronics hardware is by far less susceptible to static discharge than the electronics of yesteryears. I am still always careful around my RAM though and I always have my PC plugged in and grounded when I vacuum it. This being said, I have never had any issues from vacuuming that I am aware of. The only issues I have run into consistently over the years through my builds are PSU failures and lifespans of GPUs due to overclock. On average it seems that GPUs always burn up after a good amount of time as a result of OC compared to factory clocks and I could always resurrect the OC burn out cards via the oven trick. The cards that I did not OC never burned out. This includes my first set of Nvidia 6800 cards I had running in my first SLI build. My first big build for BF2! I never OC'd these cards and they never burned out. I still have them somewhere and they still work!. My current GPU is an Nvidia EVGA GTX 970 and it runs BF4 at max GFX flawlessly (minimal AA) @ a temp of 58c with a custom fan profile through PrecisionX EVGA standard clocks. The fans on this EVGA mod can be loud if run @ full RPM but the cards factory firmware does not even activate the fans until its temp gets up there and it never runs them @ even 50% of full RPM. When I first powered the card up - I thought something was wrong with it because the fans did not turn on. I had to game to get them to turn on. All and all a VERY nice card! I see no difference in performance running stock fan controller gaming @ 70c vs EVGA PercisionX software fan control gaming @ 40c to 58c (at standard clocks via custom fan profile). I am sure the same would apply to CPU tech. I think temperature affects longevity but has a very small if any affect on performance unless thermal tolerances are exceeded. Unless we are talking about absolute zero, super conductors and super computers - (in my experiences) running your gear at 20c vs 80c is not going to affect performance.

Posted on 2016-11-30 04:08:29
McPoopy

regarding the above.. the key to vacuuming your PC is ..'always have my PC plugged in and grounded when I vacuum it.' Germz 2016.<---
i use an air compressor at home too, but find if i spray air for too long with it, condensation builds up on my components (bad, very bad) luckily its pretty dry where i live so i just make sure i leave the door off to air out for 15 minute after.
Im not too sure about the above comment about 80c=20c performance, obviously everything throttles to protect itself nowadays but I have seen a friends computer with idle temps of 75c (due to 3 stacked GPUs and no top exhaust fans) perform very poorly

Posted on 2017-02-24 00:34:54
Nowfal X-Tase Mouktani

My CPU temperatures hit a little over 90C for a few minutes because of some motherboard settings. Now it stabilizes well under 65c at full load. could I have damaged my CPU with a few minutes at 90c?

Posted on 2015-06-08 06:14:12

It shouldn't have caused any damage - at least not if it is a modern Intel processor. They are rated to handle nearly 100C, and if they get too hot they will throttle down to protect themselves from damage.

Posted on 2015-06-08 06:49:12
JB

Thanks for the article, i recetnly bought a Metabox p650se laptop which has a i7 4720HQ i use it for 3D rendering in 3Ds Max with VRay.
I noticed that things were getting quite hot so d/l'd Speccy and Core-Temp to see what was going on. Turns out the laptop was hitting 90c on average during render! (It idols at 50c.)
I was a bit freaked out. After i placed it on a cooling pad this dropped to 85c maximum which i thought was still pretty high.

Hard to find information around the web for such a specific thing that im doing i.e: laptop 3d render core temps etc, nice to read your article and see that it should be safe as long as i don't flog it constantly.
I'm going to refrain from rendering for long periods of time at those temperatures but reckon it will be ok for test rendering.

FYI my GPU (GTX 970m) which isn't used by 3ds Max for rendering sits at 38c idol.

Posted on 2015-11-09 08:16:53
Guilherme Bianeck

In the CPU FREQUENCY BASED ON COOLING, the working tension is always the same right ?
it's just the temperature affecting the processor performance ?

Posted on 2016-03-02 19:48:07
Dr Amir

I just did an experiment recently based on my computer Intel core i3-3110 2.40GHz, that i concluded the computer should be running at average temperature. they are two separated experiments which are playing games's performance and internet performance. Based on the playing games's performance, an average temperature would be likely to perform based on article above and cooling it down may effect the performance but may increase computer's life span. Based on internet performance, it similar with playing games's performance. Sorry my bad english

Posted on 2016-04-02 22:26:34
jdlech

Thank you for the information. I find it somewhat useful. I wrote 'somewhat' because there's something else I'm far more interested in... failure rates.
The question with cooling I have is: how much does temperature affect a CPUs life cycle?
We all know CPUs flex when heated unevenly. We know extreme heating causes extreme flexing. We know flexing is bad for he CPU. But the question is, 'how bad'?
What is the average life cycle of a CPU when it is regularly heated to 50C Vs. 100C. There's naturally some variation, accounting for how uneven the heating is, the CPU type, etc.. But the question remains - is buying a liquid cooler that always keeps the CPU below 50C worth the money? Or is it just for overclockers and hobbyists?

Posted on 2016-06-09 18:35:49

That is a very tough question to answer. You could get two CPUs and simply run one at 50C and the other at 100C for years to see when they fail - but that would be a sample size of one which is not very accurate (you might have a weak CPU or something like that). You would need to have at least 10 of each before you could be relatively confident in the results (so 20 total). And by the time they start failing, they will probably be really outdated and use an old manufacturing technique so the data would no longer be relevant.

One thing I can tell you is that we configure our machines so that they don't go above ~80C when under a heavy load. Most of our customers don't put as heavy of a load on their systems as what we use to test, but over the last two years it looks like our Intel CPU failure rates in the field are about .15%. That is an insanely low failure rate - especially since we prioritize quiet operation which often results in slightly higher CPU temperatures.

Really, I would say don't worry about it unless you are overclocking. In that case, cooler temperatures probably will make a bigger difference on the longevity of the CPU. For normal operation, however, I think that you are likely to want to upgrade your system long before the CPU will likely fail even if you do let it run over 50C.

Posted on 2016-06-09 18:49:39
jdlech

Even without that information, it's good to know that temperatures over 99C can affect performance - sometimes severely. That alone is reason enough to monitor the chip temperature and alert the user at some point before then. I've always wanted something like that integrated into the OS, so even the most uninformed would be alerted to a condition that could damage the system. I've dealt with my fair share of systems with heat/dust issues to know users need it.

Posted on 2016-06-09 19:25:04
Ala Heza

My Intel i7 is what seems to heat up first before NVIDIA and It's 8 years old!

Posted on 2017-12-12 00:47:03