Stability testing IC7&PHII.

[Nemesis 1]

So whats the standard going to be. XS has same thread . The AMD guys don't want to do prime because it destroys M/B . To me that says instable. Whats a fair stability test? Anyone.

 

[SunnyD]

Why would prime destroy a motherboard? Thermal load maybe? If so, the motherboard manufacturers might want to start thinking about that issue. It might not be so much of a CPU stability issue, hence why AMD is shrugging it off.

 

[ayabe]

I've never heard of Prime nuking a mobo.

That's what I used/use Prime and Memtest.

If Prime is really killing mobos well I would have to say that product is a joke.

 

[Martimus]

Nemesis, this is the second time I have seen you posting about the IC7. What is it? I looked it up on Google, and all I could find was a ABIT MB from 5 years ago. I am sure that isn't what you are talking about. Maybe it it something to do with the upcoming Phenom 2, as you have talked about it in threads regarding that processor?

EDIT: I just remembered where I have heard about IC7. It is a diamond based TIM.

So you want to know how well the Phenom II will perform using IC7? After reading reviews of the product, it seems to perform on par with other top TIMs, but is much harder to apply. I would expect it to be about the same as other TIMs.

 

[Flipped Gazelle]

Originally posted by: Martimus
Nemesis, this is the second time I have seen you posting about the IC7. What is it? I looked it up on Google, and all I could find was a ABIT MB from 5 years ago. I am sure that isn't what you are talking about. Maybe it it something to do with the upcoming Phenom 2, as you have talked about it in threads regarding that processor?

EDIT: I just remembered where I have heard about IC7. It is a diamond based TIM.

So you want to know how well the Phenom II will perform using IC7? After reading reviews of the product, it seems to perform on par with other top TIMs, but is much harder to apply. I would expect it to be about the same as other TIMs.

IC7 = Core i7, I believe.

Originally posted by: Nemesis 1
So whats the standard going to be. XS has same thread . The AMD guys don't want to do prime because it destroys M/B . To me that says instable. Whats a fair stability test? Anyone.

Which "AMD guys" are those? My contact at AMD has no problem running Prime95 w/PhII.

 

[Nemesis 1]

I was asking what is the best stability test. Here is thread from XS. Note the guys fighting against prime 8 hours are the same people who were saying PHII was = in performance to Penryn. Penryn stability has been 8 hours prime.

http://www.xtremesystems.org/f...howthread.php?t=213248.

These are the same guys saying Buy PHII because of upgrade ability because of existing M/B. At the same time they say No 8 hours prime because of the cheap M/Bs . SO tell everone to buy because of compatability. But leave out the fact they believe The M/Bs suck. Thats what I get from the thread at XS. What I am asking is whats fair for both Intel And AMD. If AMD has M/B problem these guys say. These are true AMD fanbois. Or are the AMD M/B s fine and these guys don't want to show stability. You tell me. I have PH1. I haven't had any problems at all. But I am running stock also.

 

[AdamK47]

Any board that can be killed by Prime95 should be considered unusable.

 

[Nemesis 1]

Well heres what AMD is calling stable.

http://www.youtube.com/watch?v=wEP_s4nAZEs

Now we know thats not stable. It doesn't even belong in this thread. As were talking 24/7 use. 4 pages and not one Prime O/C shown . Even after lowering to 2 hours prime. Not looking good. FOR stability.

The reason I was asking is this. Guys will be buying these soon. Were going to see some great O/Cs . For ME . 24/7 is what I want to see . So unless I see 24/7 prime . No O/Cs will count. Its that simple . Thats the way I do Intel . I won't change rules for AMD. THe Question is how do YOU guys want to handle this.

Right now we have the Best forum on the net . As to how we interact. lets keep it that way . Come up with something we agree on.

 

[Martimus]

Originally posted by: Nemesis 1
Well heres what AMD is calling stable.

http://www.youtube.com/watch?v=wEP_s4nAZEs

Now we know thats not stable. It doesn't even belong in this thread. As were talking 24/7 use. 4 pages and not one Prime O/C shown . Even after lowering to 2 hours prime. Not looking good. FOR stability.

The reason I was asking is this. Guys will be buying these soon. Were going to see some great O/Cs . For ME . 24/7 is what I want to see . So unless I see 24/7 prime . No O/Cs will count. Its that simple . Thats the way I do Intel . I won't change rules for AMD. THe Question is how do YOU guys want to handle this.

Right now we have the Best forum on the net . As to how we interact. lets keep it that way . Come up with something we agree on.

Everyone has their own idea of stability. We have had hundreds of threads on the subject here already, and I haven't seen a single consensus on any of the threads.

I run Prime95 or Orthos Blend for 8 hours or more, and if it is stable after that; I call it stable. (I also run Memtest86 for a few hours to check the memory)

If I get problems with crashing in special applications after these tests, I lower the OC, or take it back to stock if stock speed is good enough.

If you want a consensus ATF stability test there is a sticky at the top of this forum here.

 

[Nemesis 1]

I agree with ya 100%. But lets try and setup a standard . SO we can have data on what works 24/7 for all. Anything less leads to bad threads. In that matter members requireing info about O/Cs and such have data for cpus and M/Bs with bios setting used. Seems only reasonable.

 

[bradley]

Just because the diatribes on this forum face less opposition doesn't make it the best. Moderators on other forums would have little tolerance for carrying over petty squabbles. Otherwise, I'm waiting till the AMD NDA is broken, and more than just a handful are using these chips. In the meantime, if anyone has a problem with the methodology over at Xtremesystems, sign on over there.

 

[Flipped Gazelle]

Originally posted by: Nemesis 1
Well heres what AMD is calling stable.

http://www.youtube.com/watch?v=wEP_s4nAZEs

Now we know thats not stable. It doesn't even belong in this thread. As were talking 24/7 use. 4 pages and not one Prime O/C shown . Even after lowering to 2 hours prime. Not looking good. FOR stability.

The reason I was asking is this. Guys will be buying these soon. Were going to see some great O/Cs . For ME . 24/7 is what I want to see . So unless I see 24/7 prime . No O/Cs will count. Its that simple . Thats the way I do Intel . I won't change rules for AMD. THe Question is how do YOU guys want to handle this.

Right now we have the Best forum on the net . As to how we interact. lets keep it that way . Come up with something we agree on.

AMD is not "calling" this stable... just stable enough to run Crysis. The whole thing is just a demo to stir up some excitement. It can't be taken too seriously.

I looked thru that 4 page thread on XS rather quickly, and saw 2 users claiming that Prime95 killed the mobos. They don't give any specifics on the circumstances, so to me those reports are not creditable.

If you you don't crash, lock-up, etc., then the PC is stable. Anything else - like you must run 8 hrs of Prime - is just arbitrary and therefore meaningless to those who choose not to abide.

 

[Nemesis 1]

@ bradly Well since you directed that at me . Fine. I like reading XS threads. There the most entertaining on the net. But very vicious. 3 years ago XS was about O/Cing now its about XS fanbois. The threads are entertaining but contain very little valuable tips . Other than from people like Tony and others.

After reading the thread I thought AT should do the same establish a stability test that all can agree on.

The link posted in this forum is good and the way its setup I agree. But if the AMD guys want to shorten the test time lets hash it out. I know I have run prime were errors didn't occur until 6 hours into the run. Since I don't bench for fun its not a big deal . But When I see claims of big O/Cs I want to see proof of stability. Those old P4s would boot into windows is that stable LOL. P4c would boot @ 4ghz. Very few ran their.

Thats what I am getting at . Ya see the world changed recently . Its not about performance now. Its about O/Cs stable or not. So this is a hugh issue.

 

[geokilla]

2 hours of OCCT, and like 3 hour pass of Memtest Test #5 is stable for me. NVIDIA drivers are what cause my main rig to be unstable.

PRIME shouldn't be able to destroy anything except for maybe the CPU due to high heat.....

 

[BTRY B 529th FA BN]

OCCT is suspect in my opinion, not prime

 

[dajeepster]

if I can run seti 24/7 OC with no errors or lockups for weeks... then that is stable for me.

if I get errors, i'll back off the OC til I get no errors or lockups.

 

[BlueAcolyte]

I have gotten crashes after 10 hours of OCCT.

 

[BlueBlazer]

Most of those XS guys can't get above 3.8GHz ~ 3.9GHz on air cooling... And at those frequencies they ain't 100% stable either. Sure there were 4GHz screenshots.. suicides, still not prime stable. Those voltages are crazy, even before reaching 4GHz, its already 1.5V~1.6++

 

[Markfw]

Originally posted by: Nemesis 1
I agree with ya 100%. But lets try and setup a standard . SO we can have data on what works 24/7 for all. Anything less leads to bad threads. In that matter members requireing info about O/Cs and such have data for cpus and M/Bs with bios setting used. Seems only reasonable.

Well, MY standard, is any system that can run 100% load 24/7 for at least a month. I call that stable F@H is what I use to do this.

A week or more is a good indicator.

 

[Idontcare]

Originally posted by: BlueBlazer
Most of those XS guys can't get above 3.8GHz ~ 3.9GHz on air cooling... And at those frequencies they ain't 100% stable either. Sure there were 4GHz screenshots.. suicides, still not prime stable. Those voltages are crazy, even before reaching 4GHz, its already 1.5V~1.6++

Reminds me of the early hype on yorkfield OC's. A few early-birds jumped up and down with their 4.3GHz OC's then over time more folks chimed in and the average turned out to be closer to 3.9-4GHz.

I suspect we'll be seeing the same sort of +3-sigma early results on PhII as well. Then in Feb the typical OC results (sub +3-sigma) will finally make is apparent what the reality really is/was/will-be.

 

[Cookie Monster]

If it was linpack that caused the MB death, then it would be believable but prime?!

 

[dajeepster]

Originally posted by: Idontcare

Originally posted by: BlueBlazer
Most of those XS guys can't get above 3.8GHz ~ 3.9GHz on air cooling... And at those frequencies they ain't 100% stable either. Sure there were 4GHz screenshots.. suicides, still not prime stable. Those voltages are crazy, even before reaching 4GHz, its already 1.5V~1.6++

Reminds me of the early hype on yorkfield OC's. A few early-birds jumped up and down with their 4.3GHz OC's then over time more folks chimed in and the average turned out to be closer to 3.9-4GHz.

I suspect we'll be seeing the same sort of +3-sigma early results on PhII as well. Then in Feb the typical OC results (sub +3-sigma) will finally make is apparent what the reality really is/was/will-be.

I remember that too...what a nightmare. I work part-time at a computer retailer and were constantly getting P5Ks and q6600/q9450 back stating that there was something wrong with them because they couldn't get that 4Ghz stable OC on air... bunch of noobs :|

 

[VirtualLarry]

Originally posted by: dajeepster
I remember that too...what a nightmare. I work part-time at a computer retailer and were constantly getting P5Ks and q6600/q9450 back stating that there was something wrong with them because they couldn't get that 4Ghz stable OC on air... bunch of noobs :|

You worked for a computer retailer that warrantied overclocks???

If I sold computer parts, I wouldn't even take them back if I had found out the user had overclocked them.

 

[Idontcare]

Originally posted by: VirtualLarry

Originally posted by: dajeepster
I remember that too...what a nightmare. I work part-time at a computer retailer and were constantly getting P5Ks and q6600/q9450 back stating that there was something wrong with them because they couldn't get that 4Ghz stable OC on air... bunch of noobs :|

You worked for a computer retailer that warrantied overclocks???

If I sold computer parts, I wouldn't even take them back if I had found out the user had overclocked them.

There is probably the not subtle issue of proving (as in court-of-law evidence proof) the returnee overclocked the parts as grounds for your refusal to accept them as returns. And no, a "he said to me" doesn't cut the mustard in US court when/if a customer takes you to small claims. Gets even better if they make a case of your refusing their returns because of race/gender/etc.

This country is not so business friendly on these verbal statement situations. Burden of proof falls to you (the store) to prove the customer violated the terms of the returns policy. Generally the cost of generating this documented proof exceeds the cost of the replacement the customer was wanting...so guess who gets what they want in that situation?

 

[Flipped Gazelle]

Originally posted by: Idontcare

Originally posted by: VirtualLarry

Originally posted by: dajeepster
I remember that too...what a nightmare. I work part-time at a computer retailer and were constantly getting P5Ks and q6600/q9450 back stating that there was something wrong with them because they couldn't get that 4Ghz stable OC on air... bunch of noobs :|

You worked for a computer retailer that warrantied overclocks???

If I sold computer parts, I wouldn't even take them back if I had found out the user had overclocked them.

There is probably the not subtle issue of proving (as in court-of-law evidence proof) the returnee overclocked the parts as grounds for your refusal to accept them as returns. And no, a "he said to me" doesn't cut the mustard in US court when/if a customer takes you to small claims. Gets even better if they make a case of your refusing their returns because of race/gender/etc.

This country is not so business friendly on these verbal statement situations. Burden of proof falls to you (the store) to prove the customer violated the terms of the returns policy. Generally the cost of generating this documented proof exceeds the cost of the replacement the customer was wanting...so guess who gets what they want in that situation?

Not necessarily that the parts failed, just that they were not stellar overclockers. Unfortunately, individuals who abuse the return policy in order to snag the "best" OC chip help promote the dreaded "restocking fee".

I've found that US stores, in general, are more return-friendly than other parts of the world. Associates from the UK, S. Africa and India have remarked on the greater latitude here when it comes to returning electronics, compared with the above-mentioned countries.