Burn-in works!

[Vegitto]

Out of the box, an Opteron 165 (like I have) runs at 1800 MHz. It consumes, at that moment, 1.350V. With the following method, I am rock stable at 0.976V (1.000V mobo setting).

Step 1: Download CPU Burn-in.
Step 2: Reboot at a voltage at which Prime95/StressPrime2004/Stability testing tool of your choice fails within seconds. Literally seconds.
Step 3: Load up 2 instances of CPU Burn-in.
Step 4: Set proper affinities, as the program cannot do this by itself.
Step 5: Select 'disable error checking'.
Step 6: Let both instances run for 18-48h for best result.
Step 7: Test for stability. If it's not stable, repeat the process again, for a prolonged period of time (>1.5x the original value).
Step 8: Repeat process until desired temps are achieved.
Step 9: Reap the profits of lower temps .

I can tell you, this works. I've gone from 44C under load to 38C under load, and from 35C idle to 29C idle. That's a 6C degree drop and a 6C degree drop, for free! That's better than any thermal paste can do.

My specs:
AMD Opteron 165 CCBWE 0550 UPMW
DFI LanParty Ultra-D
Stock HSF
AS5
Antec SmartPower 2.0 500W

All of the tests were conducted with an ambient temperature of 18C, in a closed-case environement.

And, this also works with overclocked frequencies. For example, if you're stable at 2800 MHz at 1.55V, after using this method, you could be stable at as little as 1.30V .

I'm currently doing this with my Opteron 165 CCBWE 0550UPMW, until I can get 2.6 GHz at stock voltage. It might not even be so far away!

Try this people. It might not work for you (although I'm really sure it will), but it can't hurt, can it? Right, that's what I was thinking .

EDIT: Thank you for pointing out that minor mistake, StumbleBum1 .

As I have forgotten to take a screenshot, I'll run dual instances of StressPrime2004 overnight and post the results.. Mind you, this'll only be 8-12h.
Stress test completed: text

 

[StumbleBum1]

Fixed Link

 

[imported_ST]

Originally posted by: Vegitto
Out of the box, an Opteron 165 (like I have) runs at 1800 MHz. It consumes, at that moment, 1.350V. With the following method, I am rock stable at 0.976V (1.000V mobo setting).

Step 1: Download CPU Burn-in.
Step 2: Reboot at a voltage at which Prime95/StressPrime2004/Stability testing tool of your choice fails within seconds. Literally seconds.
Step 3: Load up 2 instances of CPU Burn-in.
Step 4: Set proper affinities, as the program cannot do this by itself.
Step 5: Select 'disable error checking'.
Step 6: Let both instances run for 18-48h for best result.
Step 7: Test for stability. If it's not stable, repeat the process again, for a prolonged period of time (>1.5x the original value).
Step 8: Repeat process until desired temps are achieved.
Step 9: Reap the profits of lower temps .

I can tell you, this works. I've gone from 44C under load to 37C under load, and from 35C idle to 29C idle. That's a 7C degree drop and a 6C degree drop, for free! That's better than any thermal paste can do.

And, this also works with overclocked frequencies. For example, if you're stable at 2800 MHz at 1.55V, after using this method, you could be stable at as little as 1.30V .

I'm currently doing this with my Opteron 165 CCBWE 0550UPMW, until I can get 2.6 GHz at stock voltage. It might not even be so far away!

Try this people. It might not work for you (although I'm really sure it will), but it can't hurt, can it? Right, that's what I was thinking .

What HSF are you using? How long of a duration from initial install of HSF was this conducted? What were the amb temps during the course of all this testing? 0550UPMW stepping is infamous for its overclocking capabilities on lower voltage thresholds (see Lopri's 3.0GHz OC on air @ 1.45V). Lower voltage thresholds will naturally cause less heat disappation due to less current draw.

 

[Vegitto]

Originally posted by: ST

Originally posted by: Vegitto
Out of the box, an Opteron 165 (like I have) runs at 1800 MHz. It consumes, at that moment, 1.350V. With the following method, I am rock stable at 0.976V (1.000V mobo setting).

Step 1: Download CPU Burn-in.
Step 2: Reboot at a voltage at which Prime95/StressPrime2004/Stability testing tool of your choice fails within seconds. Literally seconds.
Step 3: Load up 2 instances of CPU Burn-in.
Step 4: Set proper affinities, as the program cannot do this by itself.
Step 5: Select 'disable error checking'.
Step 6: Let both instances run for 18-48h for best result.
Step 7: Test for stability. If it's not stable, repeat the process again, for a prolonged period of time (>1.5x the original value).
Step 8: Repeat process until desired temps are achieved.
Step 9: Reap the profits of lower temps .

I can tell you, this works. I've gone from 44C under load to 37C under load, and from 35C idle to 29C idle. That's a 7C degree drop and a 6C degree drop, for free! That's better than any thermal paste can do.

And, this also works with overclocked frequencies. For example, if you're stable at 2800 MHz at 1.55V, after using this method, you could be stable at as little as 1.30V .

I'm currently doing this with my Opteron 165 CCBWE 0550UPMW, until I can get 2.6 GHz at stock voltage. It might not even be so far away!

Try this people. It might not work for you (although I'm really sure it will), but it can't hurt, can it? Right, that's what I was thinking .

What HSF are you using? How long of a duration from initial install of HSF was this conducted? What were the amb temps during the course of all this testing? 0550UPMW stepping is infamous for its overclocking capabilities on lower voltage thresholds (see Lopri's 3.0GHz OC on air @ 1.45V). Lower voltage thresholds will naturally cause less heat disappation due to less current draw.

I'll edit the OP.

 

[imhungry]

Wow.

Would this affect anything short or long term?

Or would it affect OCs?

 

[buzzsaw13]

Holy crap! I'm definitely going to have to try this later, I've been dying to push my Opty 146 a little more!

 

[A5]

Edit: I need to read the whole post before responding

 

[Vegitto]

Originally posted by: imhungry
Wow.

Would this affect anything short or long term?

Or would it affect OCs?

It would affect temperatures significantley in both short and long term . In a positive way, of course .

Originally posted by: buzzsaw13
Holy crap! I'm definitely going to have to try this later, I've been dying to push my Opty 146 a little more!

I'd try that . Please report back your results .

 

[BlueWeasel]

How would something like this work on a NF3 and Sempron setup?

 

[Vegitto]

Originally posted by: BlueWeasel
How would something like this work on a NF3 and Sempron setup?

As long as your motherboard gives you options for voltage, it'll work. Watch it, with non-90nm processors YMMVEM (Your Mileage May Vary Even More).

 

[robertk2012]

Originally posted by: Vegitto
And, this also works with overclocked frequencies. For example, if you're stable at 2800 MHz at 1.55V, after using this method, you could be stable at as little as 1.30V .

Not likely. All this does is cure the

Your basically just curing the artic silver or whatever your using. I like to bump my system up over time to accomplish the same thing. If your overclocking properly you run the computer under a full load for 8-24 hours after each speed increase to make sure that its stable. This accomplishes the same thing.

 

[PlasmaBomb]

I agree with robertk2012.

Read the info on AS5 ->

Important Reminder:
Due to the unique shape and sizes of the particles in Arctic Silver 5's conductive matrix, it will take a up to 200 hours and several thermal cycles to achieve maximum particle to particle thermal conduction and for the heatsink to CPU interface to reach maximum conductivity. (This period will be longer in a system without a fan on the heatsink or with a low speed fan on the heatsink.) On systems measuring actual internal core temperatures via the CPU's internal diode, the measured temperature will often drop 2C to 5C over this "break-in" period. This break-in will occur during the normal use of the computer as long as the computer is turned off from time to time and the interface is allowed to cool to room temperature. Once the break-in is complete, the computer can be left on if desired.

 

[Matt2]

well. I've been stuck at 2.7GHZ on my Opty 170 forever.

Anything that can help is welcomed. That being said, I'm running CPU Burn-in right now at 2.8GHZ, 1.47v <---- fails prime in like 5 seconds

 

[StrangerGuy]

"Step 2: Reboot at a voltage at which Prime95/StressPrime2004/Stability testing tool of your choice fails within seconds. Literally seconds."

Chances are, if Prime95 fails within seconds, Windows will most likely crash before you get the chance to set affinities.

 

[Matt2]

Originally posted by: StrangerGuy
"Step 2: Reboot at a voltage at which Prime95/StressPrime2004/Stability testing tool of your choice fails within seconds. Literally seconds."

Chances are, if Prime95 fails within seconds, Windows will most likely crash before you get the chance to set affinities.

Not for me.

2.8GHZ 1.47v fails Prime on Core0 in literally 5 seconds. However, it's 3dmark stable and about 2 hours of BF2 stable at that speed (havent tried playing for longer than that).

 

[robertk2012]

Originally posted by: Matt2
well. I've been stuck at 2.7GHZ on my Opty 170 forever.

Anything that can help is welcomed. That being said, I'm running CPU Burn-in right now at 2.8GHZ, 1.47v <---- fails prime in like 5 seconds

What is the max FSB your motherboard can do? I assume you have tried changing your timings. What are your temps?

 

[nealh]

I have seen at several threads at XS support this

However, there is an intel engineer who says burn in is pure BS..so I tend to believe him

But if it works for you great...

 

[Matt2]

Originally posted by: robertk2012

Originally posted by: Matt2
well. I've been stuck at 2.7GHZ on my Opty 170 forever.

Anything that can help is welcomed. That being said, I'm running CPU Burn-in right now at 2.8GHZ, 1.47v <---- fails prime in like 5 seconds

What is the max FSB your motherboard can do? I assume you have tried changing your timings. What are your temps?

Max FSB is unknown on this board. The highest I've been to is 325. Right now I'm running the CPU burn-in at 311x9 (in order to keep my RAM above 200mhz)

Timings on the RAM ATM are 2.5-3-3-7. I've tried doing 3-4-4-8, but I still fail prime in under a minute at anything above 2.7GHZ, even with the voltage at 1.55v. So either this helps me or the absolute max my chip will do is 2.7ghz.

I could, however, say to hell with prime since the comp is game stable at 2.8GHZ 1.47v. FOr some reason though, it bugs me to know that I fail Prime

 

[DrMrLordX]

I tried this burn-in technique and it did not work for me. We had a thread about this months ago with many getting no results.

You're not going to hurt your CPU starving it of voltage, but you may waste a lot of time *P

 

[edub2006]

so this work or not? ehh might as well try...what do i have to loose?

so let me see if i get this.

i set it to the voltage that i can get 2.6ghz at but fails prime, then "burn in" for 18-24 hours. that should make it so i can lower the volts and still boot up at 2.6? someone explain how ths works?

 

[pelikan]

Myth.

 

[Vegitto]

Tell that to my now-stable Opteron at .976V.

 

[Jules]

Originally posted by: Vegitto
Tell that to my now-stable Opteron at .976V.

NICE

 

[Vegitto]

Originally posted by: MyStupidMouth

Originally posted by: Vegitto
Tell that to my now-stable Opteron at .976V.

NICE

I know.. On Monday, I'll try to get it stable at ~.850V.

 

[HmmmDonut]

Originally posted by: pelikan
Myth.

I don't think this is a myth. Just like everything in life some things work better for others. Same thing with taking off the IHS.