Is 2 hours of OCCT stress test ''stable OC''?

[Chris Redfield]

Hello everybody,

 

I'm currently working on overclocking my i7 10700KF and I have few questions.

1.Is 2 hours of OCCT (medium data) stable OC?

2.Should I set medium,small or large data when testing?

3.What is max temperature that I shouldn't pass with that CPU in OCCT?

4.Is there better stress test (except prime95 which is unrealistic)?

 

My cooler is AIO Arctic liquid freezer ii 360

 

Every comment is welcome.

 

[Electronics Wizardy]

Id argue no single test is enough to know if its stable, as all loads stress the cpu differently. 

 

Prime95 is fully unreastic, its using normal instructions, and not doing anything weird.

 

Id probably just use it as a main system as se iff you have weird issues.

[.Apex.]

3 minutes ago, Chris Redfield said:

Hello everybody,

 

I'm currently working on overclocking my i7 10700KF and I have few questions.

1.Is 2 hours of OCCT (medium data) stable OC?

2.Should I set medium,small or large data when testing?

3.What is max temperature that I shouldn't pass with that CPU in OCCT?

4.Is there better stress test (except prime95 which is unrealistic)?

 

My cooler is AIO Arctic liquid freezer ii 360

 

Every comment is welcome.

 

It's possibly stable, even if you had tested with small FFT and it didn't crash, there would still be a possibility for crashing in other workloads, the most difficult one to test is transient responses, it's when the load on the CPU is fluctuating even if the CPU wasn't being utilized to 100%, keep using your PC and observe if any issues arise.

[Bananasplit_00]

id run a few more than 2h, but then youre gona have to hit it with other loads like games and whatever else you want to do with it

[Mister Woof]

Personal opinion is if it passes a few hours of realbench scores appropriately in benchmarks, doesn't throw any wheas, and doesn't crash when you're using it for what you use it for, it's fine

[Chris Redfield]

11 minutes ago, Electronics Wizardy said:

Id argue no single test is enough to know if its stable, as all loads stress the cpu differently. 

 

Prime95 is fully unreastic, its using normal instructions, and not doing anything weird.

 

Id probably just use it as a main system as se iff you have weird issues.

I agree that no single test is enough.

Can you tell me what temperature is max temperature that my CPU shouldn't pass?

10 minutes ago, .Apex. said:

It's possibly stable, even if you had tested with small FFT and it didn't crash, there would still be a possibility for crashing in other workloads, the most difficult one to test is transient responses, it's when the load on the CPU is fluctuating even if the CPU wasn't being utilized to 100%, keep using your PC and observe if any issues arise.

I will run some other tests and use it daily to see if it's stable.

Should I test it with small or medium FFT?

3 minutes ago, Bananasplit_00 said:

id run a few more than 2h, but then youre gona have to hit it with other loads like games and whatever else you want to do with it

I will run it for 4 hours + some other test.

Do you have other recommendation for quality stress tests?

2 minutes ago, Mister Woof said:

Personal opinion is if it passes a few hours of realbench scores appropriately in benchmarks, doesn't throw any wheas, and doesn't crash when you're using it for what you use it for, it's fine

OK,I will run realbench as well,thanks.

[Bombastinator]

The problem is there seems to be levels of stability, and people seem to sometimes use the term “rock solid stability” when they mean “stable enough for my use case” stability.  The question become what confidence level of what stability level would a test like that produce?

[Chris Redfield]

34 minutes ago, Bombastinator said:

The problem is there seems to be levels of stability, and people seem to sometimes use the term “rock solid stability” when they mean “stable enough for my use case” stability.  The question become what confidence level of what stability level would a test like that produce?

I just want it to be stable enough that it won't crash during gaming.

 

Anyway it did crash after 45 minutes of OCCT.

However when I raised voltage and tried to run OCCT again it crashed in first few seconds.

I tried to lower frequency and rings and raised voltage again,but it keeps crashing in few seconds.

I really don't know why does it crashes so much after lowering everything.

I'm trying to achieve 5ghz with 1.300v-1350v,but it seems I can't run stress test at all now.

 

Since I'm still green when it comes to overclocking,I tried to watch few videos on youtube,but every youtuber have different methods.

[Electronics Wizardy]

52 minutes ago, Chris Redfield said:

I agree that no single test is enough.

Can you tell me what temperature is max temperature that my CPU shouldn't pass?

about 90-95c is bout the max Id run it at

[Bombastinator]

1 hour ago, Chris Redfield said:

I just want it to be stable enough that it won't crash during gaming.

 

Anyway it did crash after 45 minutes of OCCT.

However when I raised voltage and tried to run OCCT again it crashed in first few seconds.

I tried to lower frequency and rings and raised voltage again,but it keeps crashing in few seconds.

I really don't know why does it crashes so much after lowering everything.

I'm trying to achieve 5ghz with 1.300v-1350v,but it seems I can't run stress test at all now.

 

Since I'm still green when it comes to overclocking,I tried to watch few videos on youtube,but every youtuber have different methods.

When I OCed my old 4470k system, I could get it up to 4.5ghz and have it pass an hour or two of both prime 95 and fur mark.  This wouldn’t keep it crashing in games though.  4.2ghz was a crash a week in games.  Much more stable but too much for me.  4.1ghz was a crash every three weeks or so.  It took 4.0ghz for true “this doesn’t happen” stability.  It is a lousy example of the chip though. 

[Bombastinator]

1 hour ago, Electronics Wizardy said:

about 90-95c is bout the max Id run it at

This is true I think, though there is an older school opinion that if you want to slow the already slow deterioration of traces it shouldn’t be run above 75°c.  How true this is I don’t know.  Even using that old school approach though the 95°c thing should still last multiple years.  Possibly longer than the chip would even be used.  I’ve got an 18 year old chip in my basement that still works fine afaik.  Haven’t turned it on in a while.  Was it worth saving? Probably not.

Edited November 15, 2020 by Bombastinator

[Chris Redfield]

3 minutes ago, Bombastinator said:

When I OCed my old 4470k system, I could get it up to 4.5ghz and have it pass an hour or two of both prime 95 and fur mark.  This wouldn’t keep it crashing in games though.  4.2ghz was a crash a week in games.  Much more stable but too much for me.  4.1ghz was a crash every three weeks or so.  It took 4.0ghz for true “this doesn’t happen” stability.  It is a lousy example of the chip though. 

To be honest,I only want to overclock it because of better performance in games.I want to eliminate stuttering,get rid of dips and get extra FPS or two.

I'm surprised with your experience when it comes to overclocking to be honest.I know real life using of PC is ''real'' test of stability,but having it for 2 hours on prime95 and fur mark without issues and getting crashed while playing is weird.

 

[Bombastinator]

1 minute ago, Chris Redfield said:

To be honest,I only want to overclock it because of better performance in games.I want to eliminate stuttering,get rid of dips and get extra FPS or two.

I'm surprised with your experience when it comes to overclocking to be honest.I know real life using of PC is ''real'' test of stability,but having it for 2 hours on prime95 and fur mark without issues and getting crashed while playing is weird.

 

It is.  Like I said, it’s a lead chip.  I don’t know that it was even necessarily the chip I guess.  Could be a lead mobo.  Something is substandard somewhere.  All I got.

[terminalinfinity]

You need a variety of things to test OC.  Ive had things pass hours of AIDA and OCCT just to fail and have a system restart in cinebench.

[Chris Redfield]

1 minute ago, terminalinfinity said:

You need a variety of things to test OC.  Ive had things pass hours of AIDA and OCCT just to fail and have a system restart in cinebench.

I plan to test it with many different stress tests and use it for some time to see how stable it is.

[Bananasplit_00]

13 hours ago, Chris Redfield said:

I will run it for 4 hours + some other test.

 

Benchmarks are always a solid choice, always fun to see how well you score too but then its just down to playing some games or whatever and seeing if you can see any artifacts or if it craches

[Chris Redfield]

Guys,I continued with overclocking my i7 10700KF today and I have big problem with heat.

I manage to overclock it to 5ghz and I raised voltage to 1.300v,but my temperatures are reaching over 94c in first few minutes of testing.

I lowered it to 4,9ghz and I lowered voltage to 1.260v,but temps are still around 89-91c.

I cleaned my PC as well as my cooler,reapply thermal paste and I turned all fans in my PC to max,but still no progress with temperatures.

My cooler is Arctic liquid freezer ii 360

Please tell me what to do?

 

[Bombastinator]

14 minutes ago, Chris Redfield said:

Guys,I continued with overclocking my i7 10700KF today and I have big problem with heat.

I manage to overclock it to 5ghz and I raised voltage to 1.300v,but my temperatures are reaching over 94c in first few minutes of testing.

I lowered it to 4,9ghz and I lowered voltage to 1.260v,but temps are still around 89-91c.

I cleaned my PC as well as my cooler,reapply thermal paste and I turned all fans in my PC to max,but still no progress with temperatures.

My cooler is Arctic liquid freezer ii 360

Please tell me what to do?

 

So hot with overkill cooler.  The problem at that point could be an issue with a messed up cooler (unlikely but not impossible) a bad paste job (see messed up cooler) what paste are you using btw? If you grab the input and output hoses is there a noticeable temperature difference?  or the heat not being able to get out of the ihs do to processor density.  10700kf is 14nm though iirc so it shouldn’t be any worse at that than other 14nm parts.  You may have a problem not dissimilar to mine.  Your chip just doesn’t like OCs that other people are able to achieve.  I’ve heard it’s an issue with large sections of the 10 series because a 10700 is a 10900 with parts turned off.  Considering the supply issues they’ve been having it’s unlikely that the parts inside were flawless and merely turned off.  There could be at least one core, and the iGPU that were flawed which is why it’s a 10700kf and not a 10900k. You can keep running hot, look into your cooling (which may produce nothing useful), run at a lower clock, or put up with the heat. 

Edited November 17, 2020 by Bombastinator