Sen's Burn-In Technique

It's working!!!!

got an AMD 3700+ SanDiego

Max clock is 2750 (not bad from 2200 but I wants more!!!)

Right I took my vcore down to 1.599v prime stable for 5 mins. Took ages to find the sweet spot where it failed after 5 minutes.

Burnt it in for 8 hours - then primed it...

Got bored of watching prime after 23 minutes!!!!!

Took vcore down to 1.5675 and primes it - got lucky it failed after 5 minutes (previously failed after 1)

Now priming for 4 hours.

This technique is amazing.

My plan is to take vcore down to as low as it will go with prime being stable (mhz still at 2750) reckon i can get it down to 1.5v Then I'll start upping my mhz with vcore as needed to make it stable.

Does this sound OK.

Wanted 2900 - now I think 3ghz is reachable!!!

Sentinel you rock!!!

I honestally appreciate it. I feel that my silicon conditioning technique does work wonders when done properly . Glad its helping you.

What you are doing is exactally as I normally approach things. Get the vcore where you want it, then start clocking up again.

While you may not/probably not reach 3ghz you will at the least break 2.8ghz

Btw to the forums!

I have had this technique VERY VERY VERY useful. Look at my sig and see how high my winnie has got!

good guide, unfortunitly, i didnt get jack squat with mt p4 3.2e D0 maybe its my memory, since my 8KNXP board is pretty picky about the mem.

I don't know if I have been doing it right but for the past few days I have been trying to burn in my cpu.I have my pc overclocked as high as I can get it with the memory I have right now which isn't much(2.17to 2.29). I have something called fad in my background running constantly which forces my cpu to 100% and I have been using CPU BURN and SiSandra. Now I ran SiSandra on two different occasions for 3-4 hrs and CPU BURN for 4 hrs once . Now you said about your comp will become unstable. Well mine has, it keeps closing programs(SiSandra,aol,firefox,etc.) and reboots once in awhile. I can't get it to stay running longer than 5 min. Is this what you meant by unstable and if so how do you keep pressing it if you can't keep it running? I had to up my volts a quarter to let it run long enough to write this.lol
I was wondering if maybe I had my volts TOO low to do the burn in right. Maybe I should leave it where it is and try it again? I need some help!!!!LOL

JB
P.S.
Does anyone know why SiSandra would now when I try to run it as soon as it starts goes to Burn-In successful and that's it. It doesn't run the 20 or so cycles that I put it to.

This is great! I am basically already doing the same thing right now with my Rig1 @ home. I could run 200x12 stable @ 1.68v but 240x10 took v1.82v to get me prime stable, but I ran into a little instability this last friday and I decided to prime it to make sure nothing was amiss. Boink! Errored out after ~8 hours.

I brough my vcore back down to 1.68v, where it probably should have been, and prime errored almost in the same place that it did @ 1.82v. I'm now thinking that it isn't the cpu needing more voltage. I need more vdd!!! So I'll up the vdd a little and then do as Sen's guide suggests with my vcore.

Thanks Sen! I too have been looking for a how-to on your Burn-In technique. Thanks for getting us all up to speed!

Sentential said:

It works on anything made of silicon....anything!

Click to expand...

Great! My sunglasses have a silicon oxide layer. I'm burning them in as we speak!

CJ

LOL, CJ... don't go blind!

I am thinking maybe I am not going to be able to do this properly until I upgrade my memory and cpu HS and fan so that I can get a half way decent overclock out of it.
Right now I have a thermalite I believe it's called heatsink and a UV fan blowing on my cpu. I think I need to upgrade to a highperformance fan if not a better HS too.And upgrade to 1 gig of memory.

JB

question...after lowering the voltage and burning in, should i make sure my system is stable before i lower the vcore and burn in again?

futangclan said:

question...after lowering the voltage and burning in, should i make sure my system is stable before i lower the vcore and burn in again?

Click to expand...

That is entirely your call

Nice stuff.... I was erroring out in P95 after couple of minutes at 1.525 volts earlier, so I had to jump to 1.55 to get stable. I ran 12 hour burn on 1.525 volt and afterwards got 6 hours P95 stable. Bumped down to 1.50, Prime failed around 1 min, burning now at 1.50.

I don't burn-in if it's barely stable, such as your 1min prime. You should have it SLIGHTLY stable before a burn-in, but hey... it may just work out, and IF SO then let us know!

Hey, nice formula

One Q though - do this work also for the people with PhCh?
Running the burn-in program for a hour now (at least 5 more coming) 3100MHz/1.57V (not stable at all) 3200+ Venice, but at those low temps.... will it work....

Another thing - Sent, have you tried this?

Run at unstable speed @x.xxV, then instead testing for stability shave another 0.25V on the fly (through clockgen, CrystalCPUID) and continue burnin'-in.... then after another 5-6 hours shave another 0.125V/.025V until it crashes.... and then testing for stability or lower your MHz a notch and continue testing and saving another 0.125V/0.25V..... and after like 20-25 hours test what happened....
Just a tought...

This should become a trademark: "SENTENTIAL'S CPU BURN-IN"

I run burn-ins even after I've already run them enough times. I do it every time I want to stabilize a slightly stable new clock. I found 8-15 hours to be the best length. I wouldn't run a burn-in for less than that.

Because of Dread-Star's success I'm now inspired to go back and do some more burn in with my Venice. Haven't had much success on this cpu thus far but admittedly I didn't do much burn-in before.

Well, I did realize something funky about my cpu. I can run a burn-in and stabilize at let's say 295x9 at 1.525v on my cpu. It will run prime FULLY stable. Then, I'll do burn-ins at various voltages/mhz to obtain 299x9 at 1.55v prime stable. I've been prime stable at 299x9 at 1.55v for a while now, but once a week or so I BSOD from a basic thing (like opening an email). I decided to go back to 295x9 at 1.525v and that wasn't prime stable (when it actually was before). It turns out that I have to run the burn-in, yet again, to obtain stability.

So, FOR ME (I don't know how it is for others), I have to run a burn-in over again at each new clock that I want to settle with. It seems that every burn-in session only stabilizes it at that particular clock setting.

For example:
You have an a64 stable at 265x9 at 1.45v over the summer. You then run sentential's burn-in and obtain 270x9 at 1.45v fully stable (same temps, same volts). Then comes winter, and decide that since the weather is cooler you want the max OC at 1.50v instead. You run burn-ins to finally stabilize at 1.50v for 280x9. Okay, now back to the summer again and you need to back down to 1.45v again. You set it back to 270x9 just like last time, but it's not stable. The only way to get it stable again is to run the burn-in AGAIN!

Has anyone experienced anything similar to this? Now, I didn't have the chip for a year, but I was just creating an analogy to help explain what I'm trying to point out. When I shift back to an older clock setting that was already prime stable from burn-ins I still have to run the burn-in procedure again. This is so whenever I run the computer at a different setting for quite some time.

Thanks. Is that you in your sig?

Yes godm@n I have the same issue. It usually takes a burn or so to get it re-settled.

@ Dread:
Im glad your results are as good as what I normally get. Im almost on the verge of getting 4.4ghz stable myself Glad its helping

@Mollov:
While your method is interesting, I think it might be too aggressive for conditioning. It honestally takes a very long time to get anywhere.

Yeah godm@n its the same with me. I was trying to bump it up a bit from the the stats in my sig and borked the whole thing. I had to reset the cmos and everything. I'm still trying to get back to where I was voltage-wise.