The Ultra-Stable PC clubhouse

[El_Mayo]

i don't understand this group
what is it exactly?

 

[rflair]

We have entirely different processors but for mine at 3.4Ghz, it will pass prime 30 min with a BIOS vcore of 1.30625. To pass OCCT linpack 2 hours, I need 1.3370V. To pass 4 hours I need 1.34375.
In summary, I needed to pump quite a lot more volts through the CPU to make it stable for linpack.

Rflair - good effort but are you sure about the 3 years? I didn't think Q6600 had been around that long. i.e. release Jan 07, GO late 2007.

Ya was going on memory, its in the 2 range.

 

[Bundy]

i don't understand this group
what is it exactly?

Our aim is to prove that our rigs are ultra stable using the most severe tests we have available. Many of us are finding that our original OC were not that stable after all.

 

[El_Mayo]

what do you mean by "stable"?
just.. uber awesome?

 

[Studabaker]

what do you mean by "stable"?
just.. uber awesome?

No, this thread basically proves that just testing with Prime/Orthos does not make your box 100% stable under any circumstances. Overclocking 'stable' with Prime and getting crashes is one of the main reasons people say stuff like 'Vista sux it crashes all the time!' We've gotten to a whole new level of stability testing that proves for sure without a doubt that a system can handle the real maximum stress possible without a problem.

 

[sneekypeet]

No, this thread basically proves that just testing with Prime/Orthos does not make your box 100% stable under any circumstances. Overclocking 'stable' with Prime and getting crashes is one of the main reasons people say stuff like 'Vista sux it crashes all the time!' We've gotten to a whole new level of stability testing that proves for sure without a doubt that a system can handle the real maximum stress possible without a problem.

Only if you are benching both programs at the same time, is that statement true!

 

[Studabaker]

Only if you are benching both programs at the same time, is that statement true!

Why, when Linpack does the maximum possible stressing all on its own??

 

[erocker]

I've been folding and crunching (100% usage) for three days straight now. I has uber-stability. I think crunching is a great way to test stability as it goes to a good cause.

 

[sneekypeet]

Why, when Linpack does the maximum possible stressing all on its own??

Does it work the GPU too? nope, so it isnt system stability (or as you say 100% box stability), its proc, ram, and chipset stability. Loading the GPU at the same time not only makes sure the GPU clocks are solid with the CPU and the rest, it also would load the PSU to full potential. This being the "complete system stability" you are trying to achieve.

 

[Studabaker]

Does it work the GPU too? nope, so it isnt system stability (or as you say 100% box stability), its proc, ram, and chipset stability. Loading the GPU at the same time not only makes sure the GPU clocks are solid with the CPU and the rest, it also would load the PSU to full potential. This being the "complete system stability" you are trying to achieve.

Oh you meant it that way. Yeah I get that. I'm not sure if fully stressing both the CPU+GPU at the same time would make a difference in most systems though. I say this mainly because the full potential of either test is lost on the other, and because GPUs can handle insanely high temps with little to no effect on CPU temp anyway.

 

[sneekypeet]

Oh you meant it that way. Yeah I get that. I'm not sure if fully stressing both the CPU+GPU at the same time would make a difference in most systems though. I say this mainly because the full potential of either is lost on the other, and because GPUs can handle insanely high temps with little to no effect on CPU temp anyway.

I have to say I have been able to run high clocks on a card for testing with a lower OC on my proc. If I raised the proc OC I would get artifacts. I guess it wasnt so stable together now was it. As I say running both WILL make set any notions of instability aside.

I know one thing this is to me a much safer test of PSU stability than the crap application for PSU testing in OCCT. That app shut my PC down with my goods and a 950W PSU I know is good!

It isnt always about the temps!

EDIT: and I do find that I can OC and pass fur but still at those clocks get artifacts in 3D06 or say Crysis. I know neither are everyones normal routine, but I just feel loading half and half dosnt always grant stability durring a more strenuous test IMO of the newest games. Hell if it can game 2 hours of COD, GTA4, or Crysis, I wouldnt be too worried about stability anyways.

I realize this isnt an image to host, but with my PC at stock at the moment, you still get the idea.

Give me a screenie like this with an hour or so on either, and I would say its ultra stable...

 

[El_Mayo]

No, this thread basically proves that just testing with Prime/Orthos does not make your box 100% stable under any circumstances. Overclocking 'stable' with Prime and getting crashes is one of the main reasons people say stuff like 'Vista sux it crashes all the time!' We've gotten to a whole new level of stability testing that proves for sure without a doubt that a system can handle the real maximum stress possible without a problem.

still.. confused.
what makes your computers so "stable"?

 

[sneekypeet]

still.. confused.
what makes your computers so "stable"?

That the PC wont crash or blue screen while running these tests with overclocked components= Stable

 

[El_Mayo]

That the PC wont crash or blue screen while running these tests with overclocked components= Stable

how do you get such levels of stability?
magic?

 

[sneekypeet]

With the correct setting being used in the BIOS in conjunction with the correct voltages being added to each component.

No Magic, just alot of reading and getting to know your PC and all it's parts and what they "should be capable" of. Then time and some trial and error to get everything set right. No smoke, no mirror, not hot assistants.

 

[El_Mayo]

No smoke, no mirror, not hot assistants.

oh right
imagine if every overclocker did have a hot assistant though? =]

 

[sneekypeet]

oh right
imagine if every overclocker did have a hot assistant though? =]

TPU would be alot less active if that was true!

 

[Studabaker]

how do you get such levels of stability?
magic?

Well, what did I say? If your computer crashes, it's not stable. I've never had a Windows built that BSOD or locked up since like the Win98/ME days.

 

[El_Mayo]

haha
what's so great about a "stable" computer anyway?

 

[Studabaker]

haha
what's so great about a "stable" computer anyway?

It won't crash.

 

[El_Mayo]

well yeah obviously
but under normal computing
it shouldn't anyway

 

[Studabaker]

well yeah obviously
but under normal computing
it shouldn't anyway

All we're saying is,

Prime95/Orthos testing

No longer proves your overclock is stable.

That's all this is for me, it doesn't mean all-around stability to me, it just means CPU overclock stability.

 

[El_Mayo]

oh.
right.

 

[sneekypeet]

Well, what did I say? If your computer crashes, it's not stable. I've never had a Windows built that BSOD or locked up since like the Win98/ME days.

Then your not pushing it far enough. How do you tell what the MAX OC of your components are?

 

[El_Mayo]

*unsubscribes from thread*