is stress test bad for your laptop?

Hi,

I know many people do stress tests overnight for their new laptops but is that okay?
Would it hurt your laptop in anyway?

Thank you!

 

Only if it was badly designed, or if they used fairly cheap components.

 

No, it will not hurt your laptop in any way. And even if it would cause problems, your laptop has built-in safeguards to prevent anything from happening.

 

The only "hurting" it will do is wear and tear. But thats what these are meant to do. I bought my jeep to wheel, haul and drive and doing less or none of those is wasted potential just as not using your computer (within the designed parameters) in fear of hurting it.

Does that mean you should stress test and benchmark all day everyday? No that would be silly.

 

Not always, I know I tested 5 DV6000 system boards in a hotbox (103c inside, temp monitored and temp controlled). All 5 died, none of them shut off to prevent damage. 3 AMD and 2 Intel.

Also my Dell M6500 suffered a bit of an accident when my cousin was using it to play games, and blocked off the exhaust.... computer got super hot, GPU/LCD ended up dead. Diagnosed it my self. No safety sensor there.

Luckily I had complete care on that system, so they replaced GPU/LCD/all cables including top panel. no questions asked.

 

The point of stress testing your laptop is to make sure they're ok. If they fail, then simply return it since it's defective.

 

With each new system (before I fully install/customize my programs and data...) I stress test for around 36-48 hours.

I do use a Zalman notebook cooler (don't want to kill the system!) and I am very sensitive to any pauses, glitches or other anomalies that the system displays during and immediately after the stress test.

If my radar picks anything up: the system is returned for refund asap.

Of the many systems I've tested this way: I have only returned a few (less than a dozen). I did waste some time exchanging systems a few times - but my experience shows that if a symptom is exhibited by one, it will be (eventually) exhibited by all (none of the exchanged systems were ever kept - they were all part of the 'returned' ones).

When the 'as delivered' system has been tested and passed to my satisfaction: I add the parts I want to it (replace RAM, HDD, O/S, etc.) and test again for a further 12-24 hours.

At this point my data is added to the system and it is still critically monitored for the next month before I am comfortable enough to use it as a 'main/replacement' system at a client's workplace.

Overkill? Yeah.

 

A bit of stress will tell you how much a new laptop can take and what to expect in a real life scenario when you might want to encode a movie or two, edit some images, play a game, listen to some music (some of these at the same time).

Lately most of them (laptops) will use throttling to keep things going. So if it works after 24h it does not mean that much.

Good systems need to keep up performance, not just work.

It's harder and harder to find one of these (good) systems.

 

Better overkill, and exhaust all resources before calling it safe.

I do similar procedure. My poor M4600 ran for 48 hours straight on Prime95/30% gpu load. Then 100 GPU load, with only 3 workers in P95.

Any PC's we use for our work undergo:
3x Hard drive tests per drive
48 hour minimum stress test, in a row
Prime95 memory stress test, 12 hours
Finally folding@home for 8 hours+- (we sometimes don't do this).

 

I do the dip-it-in-water-for-1min test (while working), and do the impact hammer test three times.

.... none has survived so far.

But really, no need to torture overnight.

 

get a dell atg laptop or even a panasonic toughbook and see if they survive.

 

Torture testing is the best way to check to make sure your system is stable and increase the chance of finding any flaws, even loose cables, bad thermal paste job, other hardware problems. I'd rather torture test and have it fail in my return period than have it fail in six months and have to be without a laptop for a month or more for warranty repair. I'm not saying push it beyond its intended design, but push it to its maximum design potential.

 

Theoretically yes. But simply using your laptop (or any machine for that matter) we cause wear to some extent. Fortunately all devices have a usable life and in the case of the computer it will likely become obsolete before it fails completely. You should also keep in mind that computers are inherently prone to failure.

Its a catch 22. That very testing may in fact cause the failure.

Which is why that's best done at the factory and in a controlled environment. One of the best reasons you'd still want to buy from a manufacture.

 

So running running that laptop doing what it's supposed to do might cause the failure? Good, then I've done my job finding something at fault that shouldn't be there. It's one thing if you overclock like mad and do a hardware mod. But it's another if you run everything at peak speeds for an extended period, that's what it's designed to do. Nothing should cause a failure that wasn't faulty to begin with.

 

Technically a stress test pushes a system to or very near failure. Your "peak" speed/temp/stress/etc. is in fact the peak continuous operational max which is less than the total peak max before failure.

In many stems that would be 80-85% of peak max. Although I'm not sure about computers which may be higher?

 

I dont think that flies Krane. Computers peak performance is not their rated spec. Hence you can overclock them. If the rated spec truly was the peak performance, how COULD they be overclocked?

What's next using our computers for gaming is abusive because they can utilize a majority of the system to the max?

 

It's like telling people not to push the accelerator pedal too much because it might break the engine...

 

Incorrect. It saying you can't floor it and let it run all day and think you won't have issues. No machine can operate continuously like that.

 

Correction: no physical machine can operate like that - cpu's don't have that limitation.

 

Computers are not cars. They should be designed with thermals and power system in mind to run all day at 100%.

It's how servers run 24/7 365 days a year. Robust power, robust thermals.

 

As long as you don't run Furmark overnight it shouldn't damage anything.

It just depends on your preference. Some people like to test their system out for any errors and run the test for 24-48 hours while other are happy to just test for 1-12 hours.

 

that's cause you may have not tried a Toyota yet

right ... and that's why turbo boost remains on full speed all the time. Or does it ?!

hence those are called servers and not consumer grade laptops, and usually cost st-load more than what average Joe pays for a laptop.

.. Unless you have found a laptop that comes stock with redundant power supplies (i.e. two hooked up at the same time besides the battery), raid 5 drives in it, and cooling that is enough for 5 laptops using it at the same time, then I can see what you are saying ..

 

I was merely making a example of how a computer can be ran 24/7/365. Cooling, and power are key.
You need a robust cooling system (Sager-like for example).
And a good power supply system (motherboard component quality, power adapters generally are good enough quality).

Mix the two, and you can have 100% use for a long time as long as you keep thermals in check..

Servers cost a lot because of insane scalability. Massive amounts of ram, expansions, hard drive (sata controllers), multiple GPU's, etc. That's why.

 

miro gt,

not to go off topic, but Toyota is slightly above GM in real world usage and Turbo speeds do work as spec'd - still night and day difference between cars (mechanical) and cpus (electronics).

 

We don't need laptops that can run, we need laptops that can perform as intended for a fair amount of time. I'm not saying they should keep 100% load at high clocks for a week, but 30' seems more like a joke.

At least they should say "this system will run 3.1Ghz more than 30', during winter time, if used outside".

 

I don't see a laptop surviving being baked in an oven regardless of it is running or not.

When most people talk about stress testing their laptops, they are talking about leaving their laptop running stress testing programs, without blocking the exhaust ports, in a room with an ambient temperature that is conducive to human life.

You can either stress test a laptop in the normal way, like a normal human being would with their costly personal possession, or you can through your laptop in the pool or strap firecrackers to it and talk about how stress testing can damage it.

A laptop that suffers damage under normal stress testing, which is just high operating load on it's components for extended periods of time, was already defective or damaged before hand.

 

103C? LOL. Where in the manufacturers spec did you see that an ambient temperature of 103C was even vaguely OK? Most laptops manufacturers claim 35-40C max ambient operating temperature. Sure, the CPU die might run to 105C, but your 'testing' had completely predictable results. Next time also check what the max operating temp for an HDD is (hint - it is normally around 60C case temperature).

 

This wasn't a stress test. It was a test to see how far motherboards could go, surprisingly enough no safety features kicked in to prevent damage.

Read what I replied to Qing.

This is what happens when people don't read threads.

 

I can vouch for this.
My sister's Toyota Yaris has had 0 repairs since she bought it. I can't count in one hand how many Fords needed repairs the first 2 years.

 

Then why reply on the stress test thread? Talk about off topic.

And either way, putting the laptop in an environment hotter than boiling water was not a test, it was a guarantee. Finally I didn't see your reply to another forum member before I posted my observation.

 

This might at first glance have merit until you consider the likelihood of doing donuts in a 5.0 vs doing them in a Yaris. Is there any wonder its unlikely to have operational issues?

Bringing this back to computers, there are program stress tests and there are components/structural stress tests. I can remember a while back when some laptop were melting and a few even caught fire. What failed in those uses/tests? And would it matter in the end?

 

Heck even Intel offers the Extreme Tuning Utility with extended stress testing, MSI offers an overclock program, they wouldn't do this if it wasn't intended or they had concerns.

 

I am not sure, can you link me or tell me which models?

Most if it not all cases of melting components or laptops catching fire were related to:
Batteries
Bad thermal design
Faulty power components
Failure to maintain laptop (Cleaning heatsinks).

 

Toshiba Recalls Overheating Laptop Computers - AC Adapter Could Melt - Lockergnome

 

Yeah I've seen issues like that from time to time.

But what did you mean to prove by those faulty examples?

 

Same as I do with all my examples: to open you mind. You can't come to any conclusion without formed parameters. There were none.

 

I figured you meant performing a thorough test but wasn't sure.

You are right, stress testing isn't exactly thorough, but at the same time a well designed, and built laptop won't be damaged from stress testing.

Keep in mind many consumers, and enthusiast (Inc many if not most NBR members) do not have the knowledge nor test equipment to test a laptop with certainty.