Stability obsession?

I couldn't help but wondering if people are going a little over the top with the stability of their CPU's. They make it stable enough to run for days and days and days using Prime and CPU burn, etc.

Why not just set the goal of your overclock to be stable enough for what you use the computer for? If you only game in 20 minute chunks of time, what is the purpose of overclocking one's video card to be able to run 3dmark all day? Might as well get the extra clocks out of it. Same with the CPU, if you only use it for certain tasks, why not make sure it is stable enough for what you use it for? Except for distributed computing I can't really see why some people go overboard on the stability testing.

I'm a bit obsessive about stability. I don't want hardware to throw errors at me, ever. Plus, I run gentoo. When all your software is compiled from scratch, the last thing you want is a stability problem causing your software to work incorrectly, even after the hardware problem is fixed.

Gnufsh said:

I'm a bit obsessive about stability. I don't want hardware to throw errors at me, ever. Plus, I run gentoo. When all your software is compiled from scratch, the last thing you want is a stability problem causing your software to work incorrectly, even after the hardware problem is fixed.

Click to expand...

Hmmph..that I guess is a valid reason.

If your computer isn't running and doing what you want it to, then speed means nothing.

I haven't stability tested much on a count of my stock HS/F not working properly, (fixing that soon), but anyway, yeah, i game for longer periods of time (1-3 hours) and I know its just a game, but still its just nice knowing theres no chance of there being a problem where its ur computer's fault. Think about it, you get problems all the time cause of spyware/malware, viruses, OK so u fix them right? Next, ur online, something is wrong randomly and u cant figure out what it is, its good knowing it CANNOT be ur CPU or memory giving u errors...just cause its 1 less thing to consider/worry about. I keep my computer sparkling clean, and expect no errors or lags, so yeah, thats why I would stress test...and i will be soon lol. BTW i had HUGE problems w/ instablility just because of the tiniest tweak to my memory's latencys.

If a computer is unstable for a significant amount of time, in all likelihood it will eventually corrupt the hard drive. It's happened to me.

9mm and pip, very valid points, but I was not saying running an unstable overclock, just that if the computer was stable enough to do whatever the user wanted to do without any issues at all. For example, if you game, but never for 12 hours straight, and then just do internet stuff, wouldn't it be pointless to run at either a higher vcore or lower clockspeed just to get it to pass Prime and CPUBurn for a week and half?

jcw122 said:

I haven't stability tested much on a count of my stock HS/F not working properly, (fixing that soon), but anyway, yeah, i game for longer periods of time (1-3 hours) and I know its just a game, but still its just nice knowing theres no chance of there being a problem where its ur computer's fault. Think about it, you get problems all the time cause of spyware/malware, viruses, OK so u fix them right? Next, ur online, something is wrong randomly and u cant figure out what it is, its good knowing it CANNOT be ur CPU or memory giving u errors...just cause its 1 less thing to consider/worry about. I keep my computer sparkling clean, and expect no errors or lags, so yeah, thats why I would stress test...and i will be soon lol. BTW i had HUGE problems w/ instablility just because of the tiniest tweak to my memory's latencys.

Click to expand...

Well, that is true, I have had instability problems too, but I am talking about some of the extremes, like people who try to get it to pass the most stressful test that exists for incredibly insane periods of time, and just game for an hour or two and do idle workloads like surfing the web.


Just to avoid any future confusion that might possibly arise in this thread, it isn't my intent to raise an argument or debate here, just simply a "do overclockers obsess over stability" and "why do you run stability tests for much more than you would ever use your computer for" type of thing.

If doesnt work the way in which you implied. If I test my computer for 20 minutes using Prime, it doesn't mean that after 20 minutes of gaming you will get an error and crash. All that it shows is that your computer is not stable, and you are never sure when it will crash.

Put it this way, my computer is 30+ hours stable because I want the most stable platform to work on. The last thing I want is my computer to crash when I am in the middle of writing a term paper.

Let it be said again... there is no defintion of stability. Some of us prefer an hour or two of prime. I will not accept anything less than 20 or so hours. Obsessive? Maybe. I like to think of it as precaution.

Stability, like beauty, is in the eye of the beholder. Some of us prefer some atrocious beasts. To each, his own.

who wants their computer to crash in that crucial half life 2 match?

@md0Cer said:

I couldn't help but wondering if people are going a little over the top with the stability of their CPU's.

Click to expand...

Compared to what? Your view of what is stable? Over the top of what is my question to you sir.

@md0Cer said:

Why not just set the goal of your overclock to be stable enough for what you use the computer for? If you only game in 20 minute chunks of time, what is the purpose of overclocking one's video card to be able to run 3dmark all day? Might as well get the extra clocks out of it. Same with the CPU, if you only use it for certain tasks, why not make sure it is stable enough for what you use it for? Except for distributed computing I can't really see why some people go overboard on the stability testing.

Click to expand...

Can you not see that others have different feelings on how they wish to have their systems perform. I personally enjoy the testing and will work for uber stability. Why you ask? Same reason why dogs lick their nether regions. Because I Can!

My take? If the person wishes to have uber stability and comes here asking how to achieve that then that is their desire and well met I say. I will help as much as I can to get that person to their chosen path.

I think that the sticky at the top regarding those posters and their heat issues is unfair as well.

Regard the sticky titled: Sticky: Too many of you are too worried about CPU temperatures!

Again my question is Who decides how much is "Too Worried"?

http://www.ocforums.com/showthread.php?t=329015

Take heed of this thread sir. Others are telling you simply, "To Each, their Own". To misquote the Beatles...."Let Them Be".

R

ropey said:

Compared to what? Your view of what is stable? Over the top of what is my question to you sir.

Click to expand...

Over the top, as in an extreme. I should have phrased the title better and more clearly stated my point of the thread. Here as I said:

@md0Cer said:

Just to avoid any future confusion that might possibly arise in this thread, it isn't my intent to raise an argument or debate here, just simply a "do overclockers obsess over stability" and "why do you run stability tests for much more than you would ever use your computer for" type of thing.

Click to expand...

Can you not see that others have different feelings on how they wish to have their systems perform. I personally enjoy the testing and will work for uber stability. Why you ask? Same reason why dogs lick their nether regions. Because I Can!

Click to expand...

Great! My point in the thread, opinions on if or if not it is sometimes an obsession, or why some like more stability than others. Thanks for participating. Not an argument here, just a "what are your views on stability and why?" type of thing as said in the above post. Again, I apologize for being unclear when starting the thread.

My take? If the person wishes to have uber stability and comes here asking how to achieve that then that is their desire and well met I say. I will help as much as I can to get that person to their chosen path.

Click to expand...

Of course. Me too. I am sure many posts from me, you, and many other OCF members about helping a person to "their chosen path" can be found.

I think that the sticky at the top regarding those posters and their heat issues is unfair as well.

Regard the sticky titled: Sticky: Too many of you are too worried about CPU temperatures!

Again my question is Who decides how much is "Too Worried"?

http://www.ocforums.com/showthread.php?t=329015

Take heed of this thread sir. Others are telling you simply, "To Each, their Own". To misquote the Beatles...."Let Them Be".

R

Click to expand...

Thanks for the feedback, I gotta go.

@md0Cer said:

I think that the sticky at the top regarding those posters and their heat issues is unfair as well.

Regard the sticky titled: Sticky: Too many of you are too worried about CPU temperatures!

Again my question is Who decides how much is "Too Worried"?

http://www.ocforums.com/showthread.php?t=329015

Take heed of this thread sir. Others are telling you simply, "To Each, their Own". To misquote the Beatles...."Let Them Be".

R

Click to expand...

Thanks for the feedback, I gotta go.

Click to expand...

Hello! I have returned.

Anyway, that thread was created during a time when people seemed to put fear into those new to overclocking that their CPU was toast if it budged 1°C over 50 degrees.

Too worried? Can't be such a thing. Some people worry more than others. My duty to change such a thing? Perhaps not, however when someone is worrying constantly over false information, I felt I should have pointed something out.

I could have worked on that thread title a wee bit more to sound more "correct."

Also, if it was a disasterous thread, why did sevral seniors approve of it? Why did sevral moderators approve and sticky it? Why did numerous members who have been around a while approve of it and send me PM's about how good it was and wishing for more information in it ( just so no one thinks I am one cold *******, yes I happilly added more info, and some info about cooling troubleshooting as well to make it a better sticky and help more people as it has in fact, helped quite a few) ? It wasn't my intention for it to be a sticky. If anything my CPU thermal output specs and maximum casing temperature databases should be the ones stickied (although I am glad they aren't at the moment as they are not entirely finished).

If you are going to accuse me of bitching at the public with my "catch all location sticky," that was created when it was:

CPUs
--AMD
--Intel

...and "CPUs" was a "catch all" location. Moderators were working around the clock moving 50-60 percent of the threads. It wasn't a sticky then, only until the forum changes were made was it stuck.

Anyway, I really appreciate the feedback Ropey. Thanks for the opinions and have a good one!

Take heed of this thread sir. Others are telling you simply, "To Each, their Own". To misquote the Beatles...."Let Them Be".

Click to expand...

Maybe so. As the point of this thread has been mistaken probably due to tiredness and a ****ty thread title, along with the sticky you were so generous to give me your comments on.


Good advice my freind.

@md0Cer said:

I couldn't help but wondering if people are going a little over the top with the stability of their CPU's. They make it stable enough to run for days and days and days using Prime and CPU burn, etc.

Why not just set the goal of your overclock to be stable enough for what you use the computer for? If you only game in 20 minute chunks of time, what is the purpose of overclocking one's video card to be able to run 3dmark all day? Might as well get the extra clocks out of it. Same with the CPU, if you only use it for certain tasks, why not make sure it is stable enough for what you use it for? Except for distributed computing I can't really see why some people go overboard on the stability testing.

Click to expand...

I agree 100%. If your computer needs to crunch numbers for 9435 hours in a row, then sure test it in that fashion. If you're like me, i only require my systems to be stable at their max clock for 1-5 minutes at a time (benching), and my daily clocks to be stable for about a max of one-two hours of gaming at a sitting. Prime has to be the stupidest proggy in my opinion, it proves nothing to me personally. If i were to test for "24 hour stability" it would be a 24 hour bench/game loop, that way i test all aspects of what i consider to be important (CPU/Mem/GPU/Ect). Just my .02

@md0Cer said:

Maybe so. As the point of this thread has been mistaken probably due to tiredness and a ****ty thread title, along with the sticky you were so generous to give me your comments on.

Good advice my freind.

Click to expand...

Generosity had nothing to do with it really. I read the title after reading your past "stickied" thread and thought. Who is he to tell me what is "correct" feelings to have regarding "my" system?

It's all relative.

Some are not at all worried and go to the extreme of blowing their cpu's, mosfets, etc.... and still are not worried. So I say, go for it

Some are very worried relative to the extreme (see above) and others in varying degrees less so and my view is who is anyone to tell the person what their worrying state of degreeness should be?

For that reason I posted and stand to this view my friend. I do not have that right and I don't think you or anyone else does as long as the post holds to the rules of the forum.

I also feel that both threads titles are in the paragraph above's vein and just because moderators see it as sticky material still does not mean that telling others that their worries or views are not acceptable is a form of control over and above the forums rules. I also believe that the sticky needs to be deleted for this reason. It is not personal, it is a systemic patterned proof of desired control of the masses imo, no more no less imo.

R

EDIT: I just thought this would be better off in a PM, as this isn't the discussion to be going on publically as no one else obviously is getting involved.

Theres one problem i have found with running a CPU, that may not be quite as stable overclocked as it is stock. Ambient temps i found were a big problem, when i had my 2.4C, that was a few hours prime stable and would game without hic-up, but everyso often when it was really hot at night, i would get constant errors, ok my fault for not testing with very high ambient temps, but they don't come around that often. My point being that if your rig is prime stable for 3 days then some extra heat (in my case) didnt affect my CPU.

Me, i don't really do much stability wise, as long as my pc does a days work (video work games FAH) then thats good enough for me. I use prime for about an hour if it passes that, i then just set my pc to work. Once im sure its stable i load up FAH. Saying that all these things are rather CPU invensive, so maybe i am doing alot stability wise, although would I if i done less CPU intensive work , i think i probably would but 24hours prime and i'd call it quits.

I think a lot of people want there pc to be stable for whatever is thrown at it, even if most of the type its an hour of games a day then idle, or even off.

I've spoken to a few people about stability, and a few have suprised my a lot, with comments like ''I get one sudden reboot a day'' 5fps drop would probably fix this, i don't understand it heh.

I like my overclocks 100% stable, but I've found running 20 hours of prime won't tell me a darn thing, I could lock it up in 30s running other tests. I've got a little routine of my own now for stability "testing" though it's more like statistical process control than testing, takes half an hour, but it mostly relies on my own interpretation of core voltage slew and flutter under worst case load, while maintaining a very safe temp margin. Rigs that meet my approval on that test have gone on to fold for 6 weeks straight, (I usually reboot them after 2 weeks unless I forget, win98SE tends to slow down gradually) rigs that I haven't been quite sure of, and gone on to test with stressCPU or similar overnight to check, have usually locked up overnight. Anyhoo, just saying that certain repetetive tests probably don't prove much, however long you run them, if your only criteria is whether they crash or not. In my half hour test, they often won't crash, but I'll fail the rig on it, because I saw something I didn't like. I guess it's kinda statistical/empirical. I might try and have a good think about how I actually do it and how to explain it and make a thread on it sometime.

#1, running distributed computing projects, I want my results to be valid and useful for the project.

Another reason, I like my PC being 100% stable is so that if it does crash at some point, it means I can eliminate that its the hardware that caused the problem.