o/c and 'Burn in' - is it true?

Deanodarlo

Senior member
Dec 14, 2000
680
0
76
I'm currently trying to o/c a k6-III 400 to 450 and need to know if the "burn in" advice is true or a myth.

Some people claim that o/c'ing should be done in small increments, with each gain followed by running the cpu full whack for several days using benchmark programs or something similar. This effectively "burns in" the cpu at this speed making it more stable and even allowing lower voltage settings for the given speed.

Aside from cooling, is this extra o/c'ing technique true? Has anyone got experience from using this technique?

Thanks for the advice.

Deano.
 

TeMpT

Senior member
Feb 2, 2001
503
0
0
Doing o/c in small increments is probably a good thing, as there is no set rules for o/c. By doing it slowly, you are more likely to find a stable level for your box. What you want is to find out how much you can push you system while maintaining lowest voltage(thus lower temperature, which implies better stability...)

The exact time you are willing to run benchmark is up to you. Personally, I let my box rip for about 30 minutes to an hour at each stage. Once I find where my limit is, I bring it back down one notch and slowly drop the voltage until optimal. Then I let the box rip for another day to two to ensure nothing is wrong with it.

hope that helps
 

imhotepmp

Golden Member
Mar 23, 2000
1,418
0
76
Burn in is a myth, do a serach for it and youll find lots of threads about it.

Yes as tempt said, doing it in increments is a very good thing. But try not to do too small of increments if youre impatient like me. At each increment try using something htat really stresses(sp?) your system. Like rc5 <plug> Join the TA rc5 team. See the distributed computing forum for more info </plug>. Or try running multiple benchmarking software simultaneously and that will definitely stress your system.

imhotepMP
 

Justintime

Junior Member
Feb 18, 2001
1
0
0
Hiyas,
Hi all! Well as a overclocker of k6-2s for years and more recently (ok, like 2yrs now)of athlons then durons etc. i testify that burn in does indeed work! I had k6-2 500 processors that hit 550 easily, once they were burned in for abou to 2-3weeks or a few days on a full cpu stress program, with the vcore somewhere around 2.4-5!@500! LOL they easily hit 550@ 2.4 -5V ususally i find persistance pays off, and you gotta have patience, i usually give it a whippin with the vcore and cool it properly, then run it for a while, usually for a fewdays highly overclocked to maximum stable frequency, then drop the vcore and hope, though i find Athlons and durons overclock in the 100's regularly off the bat, with no real burn in needed, though with continous use, you can indeed drop the vcore a bit! all in all i find burn in towork better for the older generation (k6-2's 3+'s etc) though it works a bit on the later ones as well. Best advice i can ever give, whether overclockin or not, is to use thermal grease! whether it be arctic silver or radio shack goop, just do it! My Athlon 1.2@1.25ghz (1.74Vcore) runs in the mid 20's to about 41C MAX with fop38, i plan on getting a kt133A board soon, and reduce my multipliers etc, my Duron 800@1Ghz runs pretty well i must say! at around 30-43C with standard cooling and 1.69Vcore, i have a kt7Raid and a7pro with the duron.
BYE all, Visit me at HWC (same name), hope to be regular here as well!
 

Wiz

Diamond Member
Feb 5, 2000
6,459
16
81
As I've opined before, I believe what people call 'burn in' when they rave on about it being true and the answer to all your overclocking worries is a combination of dumb luck, settling effect, fiddling and tweaking to try to get a couple degrees cooler etc...
For example my system here:
P3 700e @ 1002mhz with 256 megs Infinion
Would run no faster than 933 with Mosel ram (before getting the Infinion - Thanks to Compuwiz1)
Temp under full load (seti@home) was 52C
Fiddled and tweaked Alpha HSF until I busted off both sides of the hold down tabs on the socket,
then I wired the HSF down with bailing wire and tweaked and fiddled with it until I could run
full load at 42C.
Mickey Mouse Jury Rigged bailing wire brought the full load temp down from 52C to 42C.
It still wouldn't run faster than about 950mhz, at that it was unstable.
Bought the infinion ram and installed a 120mm Sunon fan on the side of the case blowing in
right on the mobo around the CPU, also added a 92mm fan sucking in air at the top front of the case.
Also replaced the aging 250 watt PS with an 'AMD Athlon approved switching 300 watt power supply'
Now it runs stable at 1002mhz. Running Seti@Home, Unreal Tournament, whatever - it's like a rock.
Same CPU, faster speed. Burn in? I don't think so. Things just don't get better on their own with
time and heat and the elements. They get worse unless you put more into the mix. (Sometimes even
when you do spend more money, time and effort things still get worse). It's called entropy. To
overcome it you have to input more. It's the law.

Also, to make sure I am 'on topic' I have had several K6-2 systems. I could run them overclocked but I have no tolerance at all for instability. So while I have found these chips to be nicely overclockable, I didn't see much reason to spend much resources doing so. In terms of advantage there was very little going from a 450 to 550 or 600.
Now I give my K6-2 systems away to friends and family members who can not afford to buy a system. It makes a good little internet terminal and will word process. Putting any good video card with it is a waste of money as it gets limited by the cpu.
Just my $.02