Burn in. It works

[campy]

After getting no responses to my question in regards to my c667 not being able to go to 100 mhz fsb on asus p3b-f mb I started to get dicouraged. I have been reading that some people were in favor of burn in and others didn't seem to think it worked.
I decided to try it using sandra burn in. I upped the voltage to 1.85 and set the fsb at 66mhz and let burn in for about 15-20 minutes. Much to my surprise I rebooted at 100mhz fsb and it worked. I am now the proud father of a bouncing baby 1gig.

[DocClock aka MadClocker]

Twenty minutes eh? WOW now that's a fast burnin...most peeps burn in for a few weeks...but whatever works eh? Now that starts a whole new debate...how long is long enough?
and what volts is enough.....I think I have an idea for a survey.
Great Job there Campy, and welcome to the forum. :-)

[jeff_harrison_344]

Damn, nice overclock, congrats!

[Puma127]

Does burning in really work? WIll it hurt the chip any if i ran the voltage at like 2.0 v? Also would I have to do that through the bios or is there another way?

[surlyjoe]

you can run the voltage way up and not hurt it as long as you have GOOD cooling and watch the temps.

[Puma127]

I'm guessing when you say GOOD, you don't mean an intel retail hsf...by the way by temps are 26 c idle and 34 full load

[Spode]

my 666 did 1000 straight out the box at 1.8 volts (defualt 1.7). If I put the voltage up, it overclocks no better!! It's strange.

[logos]

gee Campy, i guess i did not see your message. 'BURN IT IN' is my standard reply to many a 'it won't go higher' query. Glad to see another testimonial and congradulations on soving the tweak!

[Puma127]

Could I burn my processor in with a intel heatsink? My temps are above

[logos]

Puma127 (Apr 13, 2001 06:01 p.m.):
Could I burn my processor in with a intel heatsink? My temps are above

yeah, just start bumping the voltage up and check the temp each time, stop when she gets hotter than you want. let her run for awhile, hours, maybe a few days, and it should cool down and then you can up the voltage a little more. i think with a cely the stock heatsink should be okay although an alpha will give you 10 or more degrees cooler if and when you need it.

mine topped out at 100-105 fsb after burn in with moderate temps (idle 30-34, running around 40) and the alpha brought that down to 19-22 idle never runs above 30

[lightning]

Wow...! you are lucky to get a nice chip! my badboy won't behave as yours! i need to burn in for an hour and let me increase for 1 mhz in fbs! stupid chip!

[batboy]

My first advice is NOT to crank the voltage up until you have better cooling. I sure don't know what Logos is talking about when he says that temps will go down in a few hours to a few days of burn in??? How??? That don't make any sense.

Idle temps mean very little. You need to be looking at CPU full load temps. Puma, are you quoting system temps or CPU temps? What motherboard do you have? I'd get more input from some of the other senior mmbers before trying something you might regret, if I were you.

[logos]

batboy (Apr 13, 2001 09:49 p.m.):
My first advice is NOT to crank the voltage up until you have better cooling. I sure don't know what Logos is talking about when he says that temps will go down in a few hours to a few days of burn in??? How???

I only know what I have seen from experience- I just figured maybe that was part of how a burn in does its magic. I usually up the voltage until the temp doesn't like it and then run benchmarks, and sure enough, it will gradually run cooler. then I up the voltage again. the box i am working on now went tortuously slow up past 1.75v, it took 3 weeks and 1.95v just to get 100fsb. now it can run at 100 back down at 1.8v.

As far as cooling goes, you just have to be carefull regardless of what you have for a hsf. your cpu might roast with an alpha or might do well the stock hsf. I have seen both. too me, that is part of the fun of overclocking: every chip is different and you can problem solve and push the envelope.

[batboy]

Well Logos, that is interesting. I'm curious if anyone else had similiar results. I check my temps on a regular basis and I've never had them go down unless I lowered the voltage or did another cooling mod of some sort. Burning in my P-III 700 did eventually help me run 1071 MHz stable at 1.9v instead of 1.95v, but I never noticed a drop in temp during the burn in (I used 1050 MHz @ 1.9v during my break in period) I always consistently check my CPU load temp by running the Unreal demo loop for 15 minutes with my Hardware Doctor running in the background. How do you measure your CPU temps? Do you mind posting what your temps were and how much they dropped?

[logos]

it has been a few months now so can't reliably say, but I remember the gist of it with this cpu...a cbo cely 633. It was with a stock heatsink and fan (i was waiting for 3 weeks for an alpha). the stock voltage is 1.65. temps ara hardware doctor on a bx133.

it went to like 83fsb at 1.65, then I started upping the voltage and didn't really get past 90fsb, so I dropped back down to 66fsb and upped the voltage until the alarm went off (i set it to 50) then backed up one and ran bench marks. I am sorry i don't remember temps, but the basic thing is sometimes after an hour, or sometimes it would take 3 days, I could raise the voltage up a notch and the alarm wouldn't go off. bx133 w/this chip tops out at 1.9v.

temps, well they ran the gamut over 50 just posting to the bios at like 1.8v to in the end idling around 32 at 1.9v.

the alpha knocked ten degrees off that to 22- then, it slowly dropped down to 19.5. i think the cpu is going because it is now back up to about 24 and it is running slower and slower, regardless of voltage. i had it at 106fsb, then 105, now it runs at 101 and the temp is going back up.

time to get a p3 700?

[batboy]

Well Logos, not to put you down, but running a Celly up to 50 degrees is probably not a good idea. My experience is that they start getting unstable about mid-40's. Running them that hot over a period of time is an invitation to disaster.

If you're getting the temps slowly sneaking up over time, it might be something simple as needing to remove the heatsink and reapplying the thermal paste. That happened to me once. Good luck.

[campy]

I may have been a little over confident. It froze on me playing a game and now I can't boot past 833mhz.
I know this chip can do 1ghz so I will keep trying. Any suggestions would be appreciated.

[Spacepiston]

I ran Sisoft burn in for about a total of 15 hours. Then I updated my BIOS on the BX133, and I can hold 900mhz, (9x100fsb) fairly stable. Temps idle around 25C and full load is 35.5C. Im using Rain as well.

Still, I am having errors. Before, the computer would freeze, and I would have to reboot. These days, I just get dropped back to the desktop. Occasionally, Internet Explorer has an error and closes, same with Yahoo! mesenger, and sometimes small errors occur (like hlpctr) when the the computer is just idling.

Ive been tinkering with I/O voltage settings. But it doesnt seem to help. I feel the 2 of 3 sticks of memory (which are low quality) may be playing a roll. I already have the system set for 3,3,3 and everything as "stable as possible".

It doesnt make a difference whether or not I boot at 864, or 900, the errors occur the same, but voltage does make a difference, something tells me 2.0 volts would seal the deal, but I havent found time to look into "wire wrapping" or "lapping"

Anyway, from my experience, BURN IN WORKS. But, it certainly does nothing within a 10 minute stretch of time. I would say the bare minimum is 5 hours to see any results.

As for CPU cooling down after burn in, I cant totally disagree with that, because my chip runs cooler now then it did before, but im a bit skeptical as to the suggested causes. I say this because the temps went down slowly, over a weeks time, and I wasnt burning in at that time. I assume its just finally settling down to the actuall temp. For reference, here are the temp differences:

Right after burn in:

29.5C Idle
37.5 Load

Weeks later:

25.5 Idle
35.5 Load

Now obviously the chip is running cooler, and I havent changed a thing.

[Puma127]

batboy (Apr 13, 2001 09:49 p.m.):
My first advice is NOT to crank the voltage up until you have better cooling. I sure don't know what Logos is talking about when he says that temps will go down in a few hours to a few days of burn in??? How??? That don't make any sense.

Idle temps mean very little. You need to be looking at CPU full load temps. Puma, are you quoting system temps or CPU temps? What motherboard do you have? I'd get more input from some of the other senior mmbers before trying something you might regret, if I were you.


I'm using an Abit SE6 mobo and using the hardware doctor to report temps...those are my CPU temps...system temp is around 24 C...Where can I get a good program to load my CPU to 100%? I usually just look at average temps when I have alot of things running (icq, winamp, ms word, etc.)

[Spode]

Spacepiston (Apr 15, 2001 11:06 a.m.):
I ran Sisoft burn in for about a total of 15 hours. Then I updated my BIOS on the BX133, and I can hold 900mhz, (9x100fsb) fairly stable. Temps idle around 25C and full load is 35.5C. Im using Rain as well.

Still, I am having errors. Before, the computer would freeze, and I would have to reboot. These days, I just get dropped back to the desktop. Occasionally, Internet Explorer has an error and closes, same with Yahoo! mesenger, and sometimes small errors occur (like hlpctr) when the the computer is just idling.

Ive been tinkering with I/O voltage settings. But it doesnt seem to help. I feel the 2 of 3 sticks of memory (which are low quality) may be playing a roll. I already have the system set for 3,3,3 and everything as "stable as possible".

It doesnt make a difference whether or not I boot at 864, or 900, the errors occur the same, but voltage does make a difference, something tells me 2.0 volts would seal the deal, but I havent found time to look into "wire wrapping" or "lapping"

Anyway, from my experience, BURN IN WORKS. But, it certainly does nothing within a 10 minute stretch of time. I would say the bare minimum is 5 hours to see any results.

As for CPU cooling down after burn in, I cant totally disagree with that, because my chip runs cooler now then it did before, but im a bit skeptical as to the suggested causes. I say this because the temps went down slowly, over a weeks time, and I wasnt burning in at that time. I assume its just finally settling down to the actuall temp. For reference, here are the temp differences:

Right after burn in:

29.5C Idle
37.5 Load

Weeks later:

25.5 Idle
35.5 Load

Now obviously the chip is running cooler, and I havent changed a thing.

apart from maybe your ambient/case temperature?

for me, my room is a different temperature each day, You have to take that into account and find a delta T.

[Phil]

I think I made a post on this topic the other day but some posts seem to have gone missing what with the servers going ons lately
anyway rather than type it twice:
http://forums2.overclockers.ws/forums/UltraBoard.cgi?action=Read&BID=1&TID=9505&SID=42324