Running LL test w/o aging my computer? - Page 2

VBCurtis · 2019-07-09, 23:52

The answer to OP's queries have more to do with which bits he's worried about wearing out. The CPU simply doesn't wear out, at least not on a timescale relevant to the useful life of the chip (say, 10 years). Static thermal loads are best for hardware; desktops that run heavy-computation tasks 24/7 are at roughly constant temperature for years at a time, never expanding nor contracting due to changes in temperature.

A throttling plan that limits frequency to keep temps and power draw low (such as a BIOS setting to not use full-frequency) is a good kind of laptop throttling; one that cycles a task on and off repeatedly to keep temp below some threshold is bad throttling, since individual parts of the CPU will alternate hot and cold thousands of times a day. That said, the package & heatsink temp will stay mostly constant, and silicon is very very unlikely to fail even under such nasty conditions.

Other parts do age faster under high-temp environments, but battery age dominates the conversation. I personally consider 5 years the typical life of a laptop, due to physical damage/battery failure/AC adapter plug failure, etc. Motherboard component failure is a risk, but for me it's a smaller risk than those other things. Under a 5-year lifetime assumption, there isn't much that distributed computing tasks will do to shorten that life (but if there's a fan, clean/de-dust the chassis yearly or more often!).

JuanTutors · 2019-07-10, 00:37

Quote:

Originally Posted by VBCurtis

The answer to OP's queries have more to do with which bits he's worried about wearing out. The CPU simply doesn't wear out, at least not on a timescale relevant to the useful life of the chip (say, 10 years). Static thermal loads are best for hardware; desktops that run heavy-computation tasks 24/7 are at roughly constant temperature for years at a time, never expanding nor contracting due to changes in temperature.

A throttling plan that limits frequency to keep temps and power draw low (such as a BIOS setting to not use full-frequency) is a good kind of laptop throttling; one that cycles a task on and off repeatedly to keep temp below some threshold is bad throttling, since individual parts of the CPU will alternate hot and cold thousands of times a day. That said, the package & heatsink temp will stay mostly constant, and silicon is very very unlikely to fail even under such nasty conditions.

Other parts do age faster under high-temp environments, but battery age dominates the conversation. I personally consider 5 years the typical life of a laptop, due to physical damage/battery failure/AC adapter plug failure, etc. Motherboard component failure is a risk, but for me it's a smaller risk than those other things. Under a 5-year lifetime assumption, there isn't much that distributed computing tasks will do to shorten that life (but if there's a fan, clean/de-dust the chassis yearly or more often!).

Thanks! I've never had a computer last more than a year running Prime95. It might be a good idea to have a sticky thread about how to take care of computer's that run Prime95, maybe even a bundled txt file with the program? I hadn't known why my computers were dying until now. Now I know I can come back to the project with my new computer that's on the way.

VBCurtis · 2019-07-10, 02:26

Wait, you know now why they died? What happened to each one? I can't tell if you're being sarcastic, as a year is a rather short life for any machine.

My best longevity was a celeron-566 overclocked to 850 that ran for ~10 years starting spring 2000. In year 8, the overclock became unstable and had to be reduced to 733ish.
The entire uptime was spent on Prime95 or LLR, 24/7.
Reasons I've had for a laptop's demise in the last ten years:
Theft
Screen connection to MB failed (flickered for a couple weeks, then died but external monitor still worked)
AC port failed (had it soldered back on, then failed again after another 18 months)
Chassis failure from drop (machine booted, but some keyboard keys failed and touchpad clicking no longer worked)

Surely my experience isn't the norm, and motherboards / other components do fail too. I even ran a $300 netbook on LLR or GMP-ECM for two years. It didn't fail, but newer machines were so much faster that I stopped spending the power on it.

kriesel · 2019-07-10, 04:01

Quote:

Originally Posted by dominicanpapi82

Thanks! I've never had a computer last more than a year running Prime95. It might be a good idea to have a sticky thread about how to take care of computer's that run Prime95, maybe even a bundled txt file with the program? I hadn't known why my computers were dying until now. Now I know I can come back to the project with my new computer that's on the way.

WOW! A year is shockingly short lifetime. Don't accept that, look into what's killing your hardware. This laptop (my main pc) is nearing 9 years old, battery wear level is 20% capacity loss, and it contains original everything internally, except I added ram at some point. And it's run prime95 throughout. (Not always 24/7, but high duty cycle). I've worn out several external mice on it. No visible display defects. Keyboard looks well worn and a few keys need a bit of extra force to register. The USB ports are loose. The clamshell hinge attach points failed so I made new brackets. The USB ports and hinges are on me, because of the way I had been handling it, it took too many dives off the couch arm to the floor. I've replaced chargers repeatedly because the cord periodically gets caught in the recliner mechanism. I have 15 and 20 year old machines that still run. They're no good for mersenne hunting because their cpus are far too slow, and don't have pcie slots for gpus, but for ordinary office type work, email or web browsing they're fine.

retina · 2019-07-10, 04:05

My eldest machine is a Dell Inspiron 500M from 2003. Banias CPU @ 1.3GHz.

Still runs fine after 16 years of 24/7 LL. But so far no primes found.

The LCD has some diagonal banding and the edges are discoloured. The non-integrated (i.e. removable) batteries have been replaced about 5 or 6 times. The 18650 cells are still easy to get and the newer ones give almost two times the runtime compared to the original ones from 2003.

The fan is still original and still runs great. I think it uses ball bearings instead of those cheap brass rings so as long as I clear the dust out of it regularly it will continue to spin happily.

The HDD is also still original, 30GB. And so far no errors. Still perfect. Show me an SSD that can claim that!

chalsall · 2019-07-10, 15:57

Quote:

Originally Posted by retina

My eldest machine is a Dell Inspiron 500M from 2003. Banias CPU @ 1.3GHz. Still runs fine after 16 years of 24/7 LL.

You got me beat! My oldest machine (still in a production role) is a Dell Precision 490 from 2006 (Xeon 5130 @ 2.00GHz). Still reliably (but slowly) doing DCs.

Which brings up an important point... Prime95/mprime is a great way to ensure kit is "sane"; rather than worrying about it degrading the hardware, instead consider it a QA tool...

CRGreathouse · 2019-07-10, 16:18

Quote:

Originally Posted by chalsall

Which brings up an important point... Prime95/mprime is a great way to ensure kit is "sane"; rather than worrying about it degrading the hardware, instead consider it a QA tool...

Yes! Otherwise you can be running failing hardware and not know it.

petrw1 · 2019-07-10, 17:43

Once I got past the old P4 Intels with the swollen capacitor issue I've never had a desktop fail running Prime95. And I always run them 24x7 at 100% load. I retire them after 5-8 years only because I can buy a new PC for less than I paid for it that is several times faster.

I did have a failed battery on a laptop but it was already 4 years old.

LaurV · 2019-07-12, 09:18

One year? Man, remove those dust clogs, and put some oil on the joints... My oldest machine still running P95 from time to time, since 2008. I have also GPUs (GTX 580) which still run, from their apparition on the market, about the same time I joined this forum...

Edit: about damaged hardware, I think I own the record with almost 20 harddisks destroyed in my life, including two hit by thunderstruck (which damaged the serial ports where the modem was connected, and the scassi board where the HDD were, but the rest of the computer, mobo, cpu, etc, survived and was still usable for years), but this has nothing to do with running P95. At that time I even didn't know about it.

Uncwilly · 2019-07-12, 13:49

Quote:

Originally Posted by LaurV

two hit by thunderstruck ..... but this has nothing to do with running P95. At that time I even didn't know about it.

The lightning was a wakeup call to get on board with Prime95. George sent it your way.

2019-07-10, 04:05	#16
retina Undefined "The unspeakable one" Jun 2006 My evil lair 1101010001001₂ Posts	My eldest machine is a Dell Inspiron 500M from 2003. Banias CPU @ 1.3GHz. Still runs fine after 16 years of 24/7 LL. But so far no primes found. The LCD has some diagonal banding and the edges are discoloured. The non-integrated (i.e. removable) batteries have been replaced about 5 or 6 times. The 18650 cells are still easy to get and the newer ones give almost two times the runtime compared to the original ones from 2003. The fan is still original and still runs great. I think it uses ball bearings instead of those cheap brass rings so as long as I clear the dust out of it regularly it will continue to spin happily. The HDD is also still original, 30GB. And so far no errors. Still perfect. Show me an SSD that can claim that! Last fiddled with by retina on 2019-07-10 at 04:08

2019-07-12, 09:18	#20
LaurV Romulan Interpreter "name field" Jun 2011 Thailand 41×251 Posts	One year? Man, remove those dust clogs, and put some oil on the joints... My oldest machine still running P95 from time to time, since 2008. I have also GPUs (GTX 580) which still run, from their apparition on the market, about the same time I joined this forum... Edit: about damaged hardware, I think I own the record with almost 20 harddisks destroyed in my life, including two hit by thunderstruck (which damaged the serial ports where the modem was connected, and the scassi board where the HDD were, but the rest of the computer, mobo, cpu, etc, survived and was still usable for years), but this has nothing to do with running P95. At that time I even didn't know about it. Last fiddled with by LaurV on 2019-07-12 at 09:22

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
adding a computer running mprime	Unregistered	Information & Answers	14	2009-02-16 14:01
how to tell if prime is running on a remote computer	crash893	Software	5	2005-09-12 13:26
Running factoring on one computer	Unregistered	PrimeNet	11	2005-07-09 04:38
running mprime on a computer I do not own	happyraul	Software	4	2004-05-06 15:54
computer seems to LAg when Prime95 is running	Firedog18	Hardware	2	2003-07-25 22:03

2019-07-09, 23:52	#12
VBCurtis "Curtis" Feb 2005 Riverside, CA 2·2,927 Posts	The answer to OP's queries have more to do with which bits he's worried about wearing out. The CPU simply doesn't wear out, at least not on a timescale relevant to the useful life of the chip (say, 10 years). Static thermal loads are best for hardware; desktops that run heavy-computation tasks 24/7 are at roughly constant temperature for years at a time, never expanding nor contracting due to changes in temperature. A throttling plan that limits frequency to keep temps and power draw low (such as a BIOS setting to not use full-frequency) is a good kind of laptop throttling; one that cycles a task on and off repeatedly to keep temp below some threshold is bad throttling, since individual parts of the CPU will alternate hot and cold thousands of times a day. That said, the package & heatsink temp will stay mostly constant, and silicon is very very unlikely to fail even under such nasty conditions. Other parts do age faster under high-temp environments, but battery age dominates the conversation. I personally consider 5 years the typical life of a laptop, due to physical damage/battery failure/AC adapter plug failure, etc. Motherboard component failure is a risk, but for me it's a smaller risk than those other things. Under a 5-year lifetime assumption, there isn't much that distributed computing tasks will do to shorten that life (but if there's a fan, clean/de-dust the chassis yearly or more often!).

2019-07-10, 02:26	#14
VBCurtis "Curtis" Feb 2005 Riverside, CA 2·2,927 Posts	Wait, you know now why they died? What happened to each one? I can't tell if you're being sarcastic, as a year is a rather short life for any machine. My best longevity was a celeron-566 overclocked to 850 that ran for ~10 years starting spring 2000. In year 8, the overclock became unstable and had to be reduced to 733ish. The entire uptime was spent on Prime95 or LLR, 24/7. Reasons I've had for a laptop's demise in the last ten years: Theft Screen connection to MB failed (flickered for a couple weeks, then died but external monitor still worked) AC port failed (had it soldered back on, then failed again after another 18 months) Chassis failure from drop (machine booted, but some keyboard keys failed and touchpad clicking no longer worked) Surely my experience isn't the norm, and motherboards / other components do fail too. I even ran a $300 netbook on LLR or GMP-ECM for two years. It didn't fail, but newer machines were so much faster that I stopped spending the power on it.

2019-07-10, 17:43	#19
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 3·5²·71 Posts	Once I got past the old P4 Intels with the swollen capacitor issue I've never had a desktop fail running Prime95. And I always run them 24x7 at 100% load. I retire them after 5-8 years only because I can buy a new PC for less than I paid for it that is several times faster. I did have a failed battery on a laptop but it was already 4 years old.