Just like the title says. In this case I am actually wanting to know if it is a CPU do I need to like take it easy for a while or punish it and then relax or you tell me!

The burn-in theory has floated around for some time and to be honest I don't think I have ever seen any real scientific method applied to proving the theory right or wrong so you got me.

For those that don't know... the "Burn in" theory is that you OC your chip and let it sit like that for a while to "burn in" and get used to the speed. Then you later increase the speed more almost as if you were started from a stock CPU speed rather than an OC'd one. I would love to see someone prove this right or wrong.

I have something...
although no scientific proof,
This 2100XP, was able to run on
1.408vCore @ 2200mhz O/C after being bombarded by 1.80-1.95vCore for the first 6mos. of it's life. I really wished I kept better records back then, and backed those file up.
This is the same processor that will boot into Windows from BIOS @ 2700+ MHZ and benched with a 1ghz OC DUT3C AIUHB 0302.
All the O/C's and results are here in the forums, I prolly could search my posts and retrieve most of it...

I don't know if it does or it doesn't.........but, I started oc'ing my rig when I first got it in November at 2228....and was able to move up on my oc to 2256........no problems.

Im gettin a new CPU within the month, but i dont know if i can test this with one rig and one CPU. Mabey 2 identical rigs and 2 CPUS. Then we could test this, im interested in if it really works too.

BTW mabey 2 people on here with similar PC's can test this. Im runnin, XP2500 mobile (the new one im gettin, dont have it yet) , COrsair XMS pc3200 512, Epox 8rda+, Radeon 9500. Mabey we should do a test, watchal think?

i have never came across any proof of burn-in, and i have been overclocking since the 386...

I believe in it, I think it does in fact work. I mean you dont have anything to loose, just give it a try. If not you havent lost anything really.

I have been pounding this DUT3C for the past 2 weeks HARD...mass FSB mass Vcore. and countless reboots..
.I shim it, and "cure" my TIM as per AS website instructs. religously.
I do have a theory on Burn IN... I belive it to be as important as breaking in a new engine. I have a scrennie that I took while IMing with Fuel ... there is a reason I think this chip runs like it does, and will operate on voltages from 1.3 to 2.19 Vcore ...Here are some screenies see what you think...

Attached Thumbnails

     

Quote: Originally Posted by Red02 I believe in it, I think it does in fact work. I mean you dont have anything to loose, just give it a try. If not you havent lost anything really.

There might be something too this especially on a new system or component, a few thermal cycles may change resistance in the curcuits, it's only speculation since none of us are engineers with the proper equipment to look at the CPU cores before and after a Burn in.

You could also do harm if you go to far and cook your CPU, Regular Athlon XPs are usually rated for 70c, and the Mobile versions are rated up to 90c. Electromigration is accelerated when High Voltage and High Temps are involved
Maybe there is a sweet spot, where the cpu gets warm enough to relax any stresses in the silicone, but not hot enough to do damage.
Red02 has seen positive results, Zennzzo is a beliver in it, hard to argue with 2.7Ghz

That does seem promising, it might be that without a burn in, the CPU would never reach the "Sweet Spot"

I would be curiuos what methods the Cryo Cooling guys on top of the ORB use, since they test countless CPU's looking for the "Magic"ones that will give them that last 10 Mhz...

There is one thing some people are overlooking when they have large OC gains after "Burn In" is the Thermal paste is settling in, Arctic Silver reccomends A Thermal Cylce is where you run your CPU at max load for a given period of time, then turn off and let cool to Ambient for a few hours.

Repeating this cycle several times allows all the Silver Crystals to align properly.
AS-5 uses several sized Crystals when aligned and filling the microscopic gaps better, allowing maximum thermal transer compared to AS3, in wich all the crystals were the same size.

There are alot of things that sound great but don't work, and some that realy do, but for different reasons than people think.
People Like Zen, who have been at it long time take these factors under consideration when saying something works, and have proof of the concept, I would think it was worth a try..

i have also noticed proof in this..i got a barton 2500, ran it at stock 3200 speeds right off the bat and required 1.88 vcore to run stable, after about a week i was able to bump down to 1.80v

Quote: Originally Posted by psychofunk Just like the title says. In this case I am actually wanting to know if it is a CPU do I need to like take it easy for a while or punish it and then relax or you tell me!

I havent really put that heory towards CPUs as such, more towards the entire rig, and the heatsink and fan.

The entire rig first off, because whenever you build a new rig, and install everything, it needs time to 'settle down' as it were. I would count this as a 'burn in method'. Its a case of the system getting used to its newly aquired flesh and blood.

The heatsink and fan explaination is very crappy but it was off the top o' my head. The paste you get on new hsf's which is already attached to the heatsink - it requires burn in time to mold around the core of the chip - athlons take presidence over P4s on this theory because of the lack of heat spreader.

As for the original overclock, I dunno - it may have to do with one or more of the chemicals used in the mixture for computer chip grade silicon. It may well be that this chemical could disperse at a higher clock speed, thus the computer runs better. im going with Solid Necros theory on resistance in the circuits - plauseable.