Sentential03's burn in's really do work for me.

[Rich Avery]

Yes on air, but it's not reliable - like now it's idling at 33'c, my max allowable temp is 55 under load but it's nowhere near that. I have the side panels off too on my case and i vented the cold air from outside which is blowing through a 5000rpm fan an inch away from my heatsink. I have a big case fan at the other end of the vent sucking the cold air down the tunnel and an old heater redirects it 90 degrees and then that 5000rpm fan taped into the heater. Obviously the heater isn't on, only the fan for the heater which is right where the 90 degree change is. Also I covered the heater in tape to stop any room air being sucked in - it's very efficient wish I had some photos.

Actually it's gone up cos I put my room heater on so I have a reliable 30'c at night with the heater off... bbrrrrrrrrrr
I'll aim for 2.7ghz with a maximum of 2.0 vcore if I can get my vcore down 0.5 I think it's doable... in fact it would probably just about run that now but it's brand new so i'm going to take it easy!!

 

[Rich Avery]

This definetely works..

I ran the GTR demo for three quaters of a lap with 1.95 in bios 1.90-1.92 abiteq before consistantly CTD or reboot, 2.0v was fine. Now with 1.95 in bios again it ran a good 10-15 laps then i exited. So time for some serious burn in time now

 

[felinusz]

deception``

There are many people who doubt this process besides myself. A debate on this particular topic could go on for hours in the same manner as one discussing the "benefits" of Prime 95. With all due respect, I am in no way out to proove everybody's theories wrong; on the contrary, I simply offered my opinion on the manner. Nevertheless, need I remind you once again that this process is not as common as it seems; therefore, I am not the only one who doubts its potential. With this in mind, please do not make it seem as if I am simply here to challenge the ideas of others as it is not the case.

Fair enough, and good points .

Something like this has no "scientific" explanation (and indeed, no explanation at all, that isn't just assumption or guesswork), and is questionable by nature. However, I still stand by saying that there's no harm in trying this particular technique, especially if there is a chance that there is some gain to be had.

deception``

Nevertheless, need I remind you once again that this process is not as common as it seems; therefore, I am not the only one who doubts its potential.

Just because the majority of the overclocking community isn't trying this, doesn't mean that they are skeptical about its value and effectiveness.

Perhaps the majority of overclockers simply haven't heard of this technique, or do not have the time to spend experimenting with it?

deception``, why not give it a shot yourself, and see if it does anything?

deception``

For most people, sentential's post was more than sufficient; for others, it still raises some questions. Obviously if his post was truly as in-depth as you claim it to be then there should be no reason for another member to make any further inquiries.

I figure that if I can get the jist of, and repeat accurately, how this technique should be performed from the original post, then the original posting was thorough enough. Maybe a five page long thread is a good reason to make further inquiries.

ok, I couldnt go through 5 pages, so I have some questions.

On another note, it seems I'm not getting any more gains out of this. After a lengthy 13 hour long CPU Burn-in session, my 2700 MHz @ 1.85V ran Prime95 for 9 minutes; which is actually down from the last attempt .

 

[enduro]

Do you restart your computer after each burn in? That can help with Prime sometimes. don't ask for a scientific explanation, but I've found that running prime after running another stress program can cause weird errors.

 

[deception``]

@fel: I have tried CPU Burn-In on previous machines, only to find that it has no performance gain. Furthermore, I no longer use Prime as a means of properly testing stability, so the technique is rather stale to me.

deception``

 

[Alacritan]

Just so you know only 12% of Americans believe that Darwin is right, and that seems like a personal attack to me. It's up to you what you believe in, if rather think that you are an evolved monkey it's up to you.

[sarcasm]

I think that if you cannot accept fundamental principals of the scientific community, the Pope of Cooling (Cathar) should excommunicate you from overclocking and revoke all computer privileges.

And we evolved from chimpanzees, not monkeys. Monkeys are not one of the Great Apes.

And besides, how can you argue with physical evidence? We've got skeletons of Australopithecus, Homo habilis, Homo erectus and Homo neanderthalensis (I know I'm skipping over a few of the less common ancestors). And most recently even Hobbits (Homo floresiensis). So did God just screw around for the last 3.5 million years up until the last 120,000 or so and decide to finally do something right and make Homo sapiens sapiens? I thought he made the world in 7 days and then plunked down Adam in a garden or something along those lines. But what do I know? I never read the Bible. I hear it has all the answers.

[/sarcasm]

 

[Alacritan]

On another note, it seems I'm not getting any more gains out of this. After a lengthy 13 hour long CPU Burn-in session, my 2700 MHz @ 1.85V ran Prime95 for 9 minutes; which is actually down from the last attempt .

It may not be your CPU holding you back at this point. Is your northbridge actively cooled? Try loosening your RAM timings and such, to reduce stress on the rest of the system.

 

[Alacritan]

Are those temps on air? If you can keep your temps in the 40's, then from the thread a while back, you should be good up to 2V or so. 2.1 might be good, but definitely make sure that your temps stay in the low 40's when you are at 2V. It's when you get toward 50°C that you start causing damage to your cpu. Personally though, I'd stay below 2V on air, even cold air. It's just safer that way.

Bartons are good to 52-55C (depends on the chip and the mobo's temp diode), regardless of how much voltage you give it. If it's got adequate cooling, you can run it as high as you want. Cathar has run his at 2.1Vcore, pretty heavily overclocked, for quite a long time now. But he's got a hell of a watercooling system. At any rate, properly cooled, Vcore is not an issue.

 

[Avg]

[sarcasm]

I think that if you cannot accept fundamental principals of the scientific community, the Pope of Cooling (Cathar) should excommunicate you from overclocking and revoke all computer privileges.

And we evolved from chimpanzees, not monkeys. Monkeys are not one of the Great Apes.

And besides, how can you argue with physical evidence? We've got skeletons of Australopithecus, Homo habilis, Homo erectus and Homo neanderthalensis (I know I'm skipping over a few of the less common ancestors). And most recently even Hobbits (Homo floresiensis). So did God just screw around for the last 3.5 million years up until the last 120,000 or so and decide to finally do something right and make Homo sapiens sapiens? I thought he made the world in 7 days and then plunked down Adam in a garden or something along those lines. But what do I know? I never read the Bible. I hear it has all the answers.

[/sarcasm]

I at no time make fun of what you belive, I'd like you to respect my beliefs as well. One of the rules at this forum is to not discuss religion. By the way there is as much evidence that proves evolution to be true as there proving it to be false. I's expect more of someone who is supposely older then I am.

 

[Super Nade]

I'm looking to lower VCore first.

Original stable settings: VCore= 1.675, CPU 257 x 10 (I''m keeping this constant for all tests)
#1 VCore=1.50 ......... No Boot
#2 VCore=1.525till 1.575 ..........No Boot
#3 VCore=1.600.....P95 4min..No Burn-in done
#4 VCore=1.6000....P95 6min...2hr burn-in
#5 VCore=1.600 P95 17 MIN after a 12 hr burn in.

Its slow and painful but seems to work. I don't know if I need to try 1 24 hr burn-in and then recheck P95 or should I up the VCore to 1.625 and run the next batch of tests. Any suggestions?

Edit***
**I ran the blend test with CPU priority set to 10. Am I wrong in running the blend test or the CPU priority?

**Enduro, I do restart after every stress test. Which makes me wonder if a short cool-down period will make any difference?

**Oops! I guess this is the reverse sentinal method I started at the bottom and now I'm pushing up VCore. Lets see how far it'll get me.

 

[enduro]

Furthermore, I no longer use Prime as a means of properly testing stability, so the technique is rather stale to me.

Doth I hear a bit of cynicism

Bartons are good to 52-55C (depends on the chip and the mobo's temp diode), regardless of how much voltage you give it. If it's got adequate cooling, you can run it as high as you want. Cathar has run his at 2.1Vcore, pretty heavily overclocked, for quite a long time now. But he's got a hell of a watercooling system. At any rate, properly cooled, Vcore is not an issue.

Sorry m8, but it's been proven that if you can't keep your temps in the low 40's when you are over 2V, you will damage your cpu. Maybe not kill, but damage it. The simple fact is, air cooling can't cool the processor down to low 40's full load with 2V, because even WC setups have trouble with it. Leave the +2V to the WC and phase change arena. No point in damaging your cpu and/or killing your overclock or cpu.

I believe Super Nade was reading the same thread about this. Hopefully he can back up this statement (psstt. SN, little back up here )

 

[Super Nade]

lol ^^. I'm on the same page as Enduro. Damage to the CPU is a very good possibility. Just ask the beater of all CPU's JP. I believe he killed his CPU by letting temps do their own thing.

Look here for the discussion(s):-
http://www.ocforums.com/showthread.php?t=337499
This sums it up:

Yeah, just overclocking it doesn't kill the cpu. It's the resultant heat in combination with high voltage that will kill the processor.

and this one started by JP:
http://www.ocforums.com/showthread.php?t=334435
Read the post #1 (with the link..READ in particular.)

 

[pelikan]

I would like to state my opinion on doing a burn in to increase an overclock. This is just my opinion and I don't mean any offense. I've done it quite a few times because I used to believe in it. I don't think that burning in changes anything in the cpu. What I think really happens is you try a ton of different settings while burning in and so have a greater chance of finding a faster setting that works.
I'm sure most of you know this, but the definition of "burn in" has to do with testing electronics for reliability, not trying to increase an overclock.
Google: define: burn in:
A given period of time in which a system is fully stressed to make sure there are no faults.
The process of operating a power supply (usually at full load), typically in an elevated ambient temperature, immediately after manufacture. This process is useful in eliminating early life failures.
A method used to screen out a component with early life failures.

 

[enduro]

Say waaa? Get out of with that stuff jking

I'm not sure if I'd agree with you on that. I've tried almost every setting in BIOS for upping fsb and core clock. What I'm doing right now is lowering the vcore until Prime will only run for a few minutes, and then stressing the cpu for hours. It's nothing that I haven't tried before, but usually I'd only do it for 30 min's tops, so no results were found. After doing it long term, the results have come back, and there's no denying them. I know it sounds weird because it shouldn't affect the cpu, but for some reason it does.

 

[Super Nade]

Then how would you explain my results (modest if I may add)?
Why should the P95 time increase? I changed no other setting at VCore=1.6 V except the stress test time.

 

[pelikan]

Yeah, you guys are experienced overclockers and I am not denying the validity of your results. If it works for you then great.

 

[Super Nade]

pelikan:
we are not trying to badger you . The question you raised is a valid one in context of your experience. But quite a few (a majority maybe?) of overclockers have reported better results after a burn-in. True, some of them seem to have achieved no noticable gain. It is very hard to say who is absolutely right since people tend to see their answer in the context of their experiments. I'm for a burn in simply because I've heard more people report favourable results. Nobody is saying you are totally wrong.

 

[subtotal]

just thought i'd[pst my results, did this a little while ago & didn't really know what i was doing (i used sandra burning wizard) but it got me an extra 200MHz at the same Vcore, i think i'll have another go at it using this method on the weekend

 

[Sentential]

just thought i'd[pst my results, did this a little while ago & didn't really know what i was doing (i used sandra burning wizard) but it got me an extra 200MHz at the same Vcore, i think i'll have another go at it using this method on the weekend

Keep the resuts comin people. If I get enough, ill start another thread to be sticked

 

[felinusz]

deception ' '

Furthermore, I no longer use Prime as a means of properly testing stability, so the technique is rather stale to me.

May I ask what you *do* use for stability testing your processor? Prime95 is just about as good as it gets.