How Realistic is "Burn-in"?

B

blade52x

2[H]4U
2FA
Joined
Sep 9, 2006
Messages
3,232
I was hoping for 3.2Ghz stable with my system on stock VCore of 1.325V, but it failed 4 hours and 15 minutes into Go-Orthos dual core blend test.

If I keep stressing the CPU at these speeds without touching the VCore, is it realistic that it could actually become stable? And if so, how long would constantly stressing it take? I just want to know so I don't waste my time on something like this. I could just slightly bump the VCore, but I am searching for my max stock VCore overclock. :)
 
Burn in to get a better overclock is mostly a myth. I have never gotten any higher after "burning in" than I did with the original careful and meticulous overclocking sessions.....my cpu's typically get overclocked and settled into their max stable speed within a day or two.
 
blade52x said:
I was hoping for 3.2Ghz stable with my system on stock VCore of 1.325V, but it failed 4 hours and 15 minutes into Go-Orthos dual core blend test.

If I keep stressing the CPU at these speeds without touching the VCore, is it realistic that it could actually become stable? And if so, how long would constantly stressing it take? I just want to know so I don't waste my time on something like this. I could just slightly bump the VCore, but I am searching for my max stock VCore overclock. :)
Click to expand...


You have a 1% chance at succeeding with that task.

You found it, Now add some volts and make my X2 4400 @ 2.650 ghz look really slow.
 
The-One said:
[/COLOR]

You have a 1% chance at succeeding with that task.

You found it, Now add some volts and make my X2 4400 @ 2.650 ghz look really slow.
Click to expand...

Well I'm testing 3.15Ghz now (350x9) and it's over 4 hours and 30 minutes so this may be my max stable on default VCore :p If it passes 8 hours, I'll drop it to 3.1Ghz and pretty much be gauranteed stable - my RAM already passed like 9 hours of memtest86+ last night at 875Mhz 4-4-4-10 so I think I'm almost ok there (just need this blend test to pass)

And am I not overvolting my CPU, so how can I "burn it out" - please tell me that can't happen with stock VCore with an unstable overclock :confused: :confused:
 
Big reason for doing a burn in is to get the thermal grease settled(artic silver for instance will drop a degree or 2 after its burn in period) and to test for system stability. It really will not increase your ability to overclock. As far as not touching the voltage it is still possible to damage something. Just not as possible.
 
swatbat said:
Big reason for doing a burn in is to get the thermal grease settled(artic silver for instance will drop a degree or 2 after its burn in period) and to test for system stability. It really will not increase your ability to overclock. As far as not touching the voltage it is still possible to damage something. Just not as possible.
Click to expand...

How unstable would you have to be risk something like that? Running at 3.2Ghz I failed 4 hours and 15 minutes into the test - doesn't seem to bad, such as other overclocks which might fail 5-10 minutes in. Now at 3.15Ghz I've been stressing 5 hours and 7 minutes and still going.
 
blade52x said:
How unstable would you have to be risk something like that? Running at 3.2Ghz I failed 4 hours and 15 minutes into the test - doesn't seem to bad, such as other overclocks which might fail 5-10 minutes in. Now at 3.15Ghz I've been stressing 5 hours and 7 minutes and still going.
Click to expand...

To me it doesn't matter how long it takes an overlclock to fail if it fails. My system runs 24/7 so I may define stability different then you. Personaly I would do a burn in test for 48 hours at 100% load before I would even think of leaving it at that setting. Most lockups would happen by then.
 
I do a 24 hr run of Orthos with a loop of RTBDRBL to fully load both the CPU/GPU and PSU.

that REALLY seems to weed out stability issues ;)
 
Arcygenical said:
I do a 24 hr run of Orthos with a loop of RTBDRBL to fully load both the CPU/GPU and PSU.

that REALLY seems to weed out stability issues ;)
Click to expand...

CPU & GPU? Can Orthos do that? :p

Anyway, up 7 hours 16 minutes stable, I'll keep this one going for 24 hours, I think I want to keep it at 3.15Ghz, but I have to be sure first (I might wait 48 hours - depends how busy I am tomorrow)
 
I conisder 4 hours stable, as long as no other programs crash then I don't care if it can run 2 days of torture testing because I'll never need to do that in the real world. Even if you do notice a little instability, the worst that could happen is a computer crash every few months. I would keep it at 3.2 and stop obsessing over stability.
 
Burn in does nothing, only checks for instability, it does NOT increase your overclocking potential.
 
I don't even care about >2 hours failures, but I am also a fast and dirty overclocker. I'm the king of "Do as I say (don't overclock like I do) not as I do" in that category... My Opteron fails Prime95 2 hours in. Yet I have not had a game lockup, 3DMark freeze, random restart, blue screen or anything for at least since I got my Raptor replaced (3 months ago). As long as I can do what I want with my PC without issue, I don't care what Prime says.
 
Sovereign said:
I don't even care about >2 hours failures, but I am also a fast and dirty overclocker. I'm the king of "Do as I say (don't overclock like I do) not as I do" in that category... My Opteron fails Prime95 2 hours in. Yet I have not had a game lockup, 3DMark freeze, random restart, blue screen or anything for at least since I got my Raptor replaced (3 months ago). As long as I can do what I want with my PC without issue, I don't care what Prime says.
Click to expand...

Well I want to be sure of stability, that way if something does go wrong, it probably won't be a hardware failure.

8 hours and 26 minutes stable and still going. This overclock seems promising and pretty safe so far (at 400x8 I was getting odd voltage spikes once in a while from ATI Tool - no idea, but I haven't ever seen them ever except for on this settings, so I think I'll just keep using the x9 multiplier)
 
I tried burning my 2800+ sempron64 in at 2.76ghz 1.65v because it wouldn't boot at 2.8ghz. After little over 2 days of prime full load I rebooted and tried 2.8 and what do you know it did nothing. I will never try burning in again because it's a bunch of crap.
 
Burning in is mainly to test stability rather than to get a CPU used to overclocking. Burning in to get higher speeds is a bunch of bullshit.
 
swatbat said:
Big reason for doing a burn in is to get the thermal grease settled(artic silver for instance will drop a degree or 2 after its burn in period) and to test for system stability. It really will not increase your ability to overclock. As far as not touching the voltage it is still possible to damage something. Just not as possible.
Click to expand...
Actually, Arctic Silver requires several full thermal cycles to "burn in." So, it's not just about getting hot, it's also about cooling back to room temperature. . . then heating up again. . . and so on.

I doubt it makes much of a difference, however, whether the peak is at OC or non-OCed temps.

H
 
Ok I'll let this run for another hour (will have completed 10+ hours by then), then I'll run super PI and do some benchmarks, and leave it at this.

I do run F@H but it doesn't stress my computer as much as Orthos does. The temps are 4-5C lower when running F@H.
 
I think burning in can work, I had my Opteron (2.7GHz) at 1.55V because it would crash prime in 20 minutes at 1.525V, after a week of using it and not likeing the heat it made, I lowered it to 1.525V again to try it out, and it lasted five hours. My MX-1 was already cured because I had been using without overclocking the processor for a couple months. I could also raise the clocks if I kept it at 1.55V but only a little.
 
firewolf said:
I think burning in can work, I had my Opteron (2.7GHz) at 1.5V because it would crash prime in 20 minutes at 1.525V, after a week of using it and not likeing the heat it made, I lowered it to 1.525V again to try it out, and it lasted five hours. My MX-1 was already cured because I had been using without overclocking the processor for a couple months. I could also raise the clocks if I kept it at 1.5V but only a little.
Click to expand...

What were ambient temps?
 
ryan_975 said:
What were ambient temps?
Click to expand...
Between 30 and 32 degrees, my core Idles around the ambient and loads about 8-10 degrees above that. That's not the reason though, I've had my processor running this way for 8 months.

I meant 1.55 versus 1.525, not 1.5 vs 1.525, sorry, I think you go the idea though.
 
You must log in or register to reply here.
Back
Top