Do computer hardware require "burn-in"?

Haha, hi everyone.

So I had that random question come up during work... so I kinda had to ask. lol

So do hardware require burn-in so that they can work work optimally? Especially for CPUs and GPUs?

If so, how can one "burn-in" their computer hardware? I'm assuming just run benchmarks (like furmark) for a long time...

Thanks for answering my utmost random question. lol

 

No. Fully stressing new hardware is sometimes done to make sure that it works properly under all circumstances, but it's not required.

 

If you overclock, then it's a good idea to give it a good run-in, otherwise they're already (or should be) tested at the component and assembled level.

 

I don't use the term burn in.
It is very misleading on its real purpose.
What it is supposed to do is to stress the system to redline, if system operates normally at redline you can be sure it should have zero issues under normal conditions.

But many people like to equate burn in as something to do like cars to make sure the fuel/lubricant "flows" into the system optimally which is wrong.

Computers don't work anything like cars, there is no fluid mechanics involved.

Intel/AMD/Nvidia already stressed tested the CPU/GPU chip to operate/compute correctly at standard operating conditions.
Compulsary Burn In is a myth.

 

you dont need to burn in computer made in the last decade. however testing to see if it is dead on arrival is crucial.

 

No, not required but highly recommended during the return period of wherever you bought it from. Don't want to find out later down the line when you are compiling that you get freezing/BSOD.

 

Right. First thing I do after playing with it for a day or two is do some Furmark, Prime95, and MemTest86+ tests for extended periods. Cycle some 3DMark, and play taxing games like Bad Company 2 (or whatever is most taxing at the time I get the laptop).

If it survives my first week with it (I really push it) then it's usually ok. Obviously monitor temperatures and look for any signs of instability too like texture tearing, etc, somethings that just don't show up as errors or crash the computer.

 

Want to try something?

Try simply installing (from the hidden HDD partition) the O/S two or more times. The second/third/fourth time (depends on the notebook, in my experience) will be a noticeably smoother running system.

Burn-in? Not a myth: a requirement.

(Try it with music/sound equipment, TV/video equipment - any electronic equipment... a few 100's hrs burn in does wonders).

 

Stress testing should be done to make sure the chip is in working condition.

but Burning it in so that it performs better or more optimally than the next chip, no such thing.

 

This is interesting. Source?

 

IMO no such thing as burn-in.

You can stress test if you are running outside of normal parameters (overclocking) to make sure it is otherwise stable.

Heck, you almost don't need much of a "break in" period in new cars these days, due to tighter manufacturing tolerances.

Just drive it like you stole it!

 

Try it with a handcarved upright bass, for about 100 years.

Source, jazz musicians of WPUNJ

Seriously, its widely considered to be fact. Although, you can also get a 10,000$ bass that sounds really good and was made in the last century...
I guess its a little different with laptops

 

Acoustic instruments need to be seasoned based on its geographic location no doubt. Strange things happen to them if you don't take care of them especially a new grand piano.

[sarcasm]Electronics is similar because electronic components' characteristics change over time. For example, my TV is about 10 years old, and now it's starting to flicker whenever there's bright colors. Now that's what I call "burn-in"[/sarcasm]. Seriously though, they'll degrade over time. So if you want to see if an electronic will improve in the next 10 years due to "burn-in", it's time to move on to something better. Electronic components undergo stress tests to ensure they are within specifications. We're talking about objective numbers here; Actual statistical data. In other words, they've done the burn-in (stress test) for you already.

I had to look up the word "burn-in" to be sure I'm not confusing it with what audiophiles use it as (improvement over time when subjected to stress). It's another word for stress test. The intent is to check the failure rate. If it fails, it fails. If it works within specs, it works. Nowhere did it mention "improvement." So if you want to see if the CPU becomes faster and faster or the GPU graphics performance improves with "burn-in" (audiophile's definition), then I'm afraid you'll be disappointed.

If you OC it, the skies the limit; move outside its recommended specs at your own risk. You can OC till you kill it or shorten its lifespan.

 

You guys are sorely mistaken if you think each and every component is 'burned in' before being sold. I'd be surprised if even 1% is tested (and, I'm sure they would be tested till they died - they wouldn't be put back into the retail channel).

You don't have to believe, but the 'test' I gave is easy enough to try for yourselves.

 

Tiller, while I usually respect your knowledge and experience and opinion, in this case I can't and don't believe it. Installing the OS multiple times will NOT make a difference. "Burn-In" as mentioned which is really just stressing your system to its limit, is a very good idea to make sure your system is stable within normal parameters. Otherwise multiple installations won't help matters unless you're using a Momentus XT drive. But even then, it wipes the NAND with a hard drive format anyhow so probably wouldn't help much.

 

No problem, like I mentioned, you don't have to believe that burn-in is beneficial. I have mentioned multiple times over the last two years on this forum that I can feel very minute differences on my equipment.

On my camera's: they definitely feel better after about 20K shutter actuations. On my sound equipment; after a month (at least) the sound is less 'mechanical' and more transparent (I hear the performance, not the equipment trying to reproduce that performance). With my cars; I break them in as if they're each limited edition Ferrari's - several friends/acquaintances who believed that they bought cars that were already 'broken in' have replaced major parts (if not the whole engine, essentially) because they 'drove them like they stole them' from day one. (They now think that the dealer/salespeople said it on purpose to sell them a car or at least parts, sooner).

However, do you really believe that every computer component/system that is shipped has been burned-in/stress-tested?

I'm surprised how many people refuse to believe how much difference this makes - no worries: but when they ask why their identical hardware is not performing at the same level as mine (sometimes, I even have inferior hardware, yet it performs better than the latest and greatest an acquaintance may be showing off) - I simply tell them that my equipment has the 'tiller touch.

Because, just like you, they would be hard to convince that running a new system for a week or two straight would make any difference on how it runs for the next 3 or 4 years.

 

I'm skeptical as well. Do you have any proof or evidence?

I can understand "burn in" on mechanical or analog systems-- In the case of your camera it could be the mechanism holding the shutter getting worn down so that the shutter moves more freely and in the case of the headphones it could be the vibrations from the sound physically affecting the driver.

However, I don't see this happening with computer chips. Transistors will operate at whatever the clock frequency is set at. Even if they somehow do get faster after "burn in", this would never show because it would still operate at the same clock speed. Think of it this way: a higher binned cpu operating that the same clock speed as a lower binned one would not be faster, would it?

 

Thanks everyone for your inputs.

Yeah... I asked this question because I'm quite the audiophile myself and I let my headphone speakers, stereo speakers "burn-in" for a long time.

I guess the main difference between those type of hardware versus computer hardware is that speakers have moving parts, the subwoofer, so letting those moving parts to move freely with little to no strain makes sense.... but since computer hardware doesn't have any moving parts, except for those mechanical hard drives, it shouldn't make sense to "burn" them in. lol

But who knows? just like what tiller said... maybe it could make a difference on how it runs for the next 3 or 4 years.

Well again, thanks everyone for your input much appreciated.

 

In the past car engines get a little better after the first 10000km (6000miles) when driven in a gentle manner. This doesn't happen any more since the burn-in is done in the factory. A car engine will deteriorate if the engine is stressed too much at the beginning of its life as mechanical parts need a bit of time to adjust.

There was a rummer in the overclocking community that stressing computer part will help them overclock better...just a rummer though.

Electronic component are like gates through with electrons pass. The speed on a component is defined by the number of electrons that can pass through a gate. Moving electrons don't have the ability to make the gate larger, hence the number of electrons passing through the gate is constant during the life time of the component, and hence the performance remain constant until the component fails. Whether a component is stressed or not its performance should remain the same.

You said that the performance of an HDD increases after re-installing and OS a few times. If what you're saying is true the read/write speed of an HDD should increase with time. I don't do that as often as before, but during a period of time I used to format my harddrive and re-install windows every 2-3 months, and tested my harddrive with sandra sisoft after each install. I didn't notice an improvement in the performance.

I'm not clear about the numbers, but from each batch of components a sample is taken and stress-tested. There is no point in testing every component a factoring makes as all the components created during the research and development phase are thoroughly tested.

One's computer does not perform like yours because not everyone is a computer geek (geek in the good sense).
If you're like me the first thing you do when you get a new computer is to re-install a fresh copy of windows. You do this to get rid of all the blot-were. every now and then you run a registry cleaner and repair tool and a defrag tool. Where ever you feel your computer is getting slow, you format and re-install windows. Last, if you want to test a new software, you don't use your main computer but a second one or a virtual machine with a fixed harddrive size.

 

I have sometimes find that a burn-in is required for whatever thermal paste manufacturers use. In a few notebooks I have owned the temperatures dropped significantly, like 5C+, after stressing the components.

 

I dunno. With desktop cpu's on s775, some of my dozens of chips would overclock slightly better after a "burn in", not to mention that thermal paste at the time often needed a burn in time to optimally cool a chip.

 

Some thermal past like AS5 needs some curing time. AS5 for instance needs about 240 hours to reach its optimum performance. The curing is done under normal condition and thus no stressing is needed. Other thermal pasts such as MX-3 doesn't require any curing time and thus the performance doesn't improve over time