Do CPU burn ins work?

[boyRacer]

Does it really increase overclocking potential? What do you use to burn it in anyway?

 

[Rand]

Whether it "works" is a matter of debate. though personally I don't bother and I very much doubt it in any way impacts potential overclockability.
Besides AMD/Intel has put each and every processor die under infinitely more stress and a far more well rounded and harsh testing scenario prior to releasing it then we could ever do.

 

[WhoDeeny]

<< Whether it "works" is a matter of debate. though personally I don't bother and I very much doubt it in any way impacts potential overclockability.
Besides AMD/Intel has put each and every processor die under infinitely more stress and a far more well rounded and harsh testing scenario prior to releasing it then we could ever do. >>



This is not true, these guys have been making these chips for so long and since the manufacturing process remains the same for the chips they gave up that process. to improve turn around time they don't do testing like that any more and if we're lucky they take one out of every thousand and stress test them, if we're lucky. So chances are high that that's the first time your proc's being used...

 

[Rand]

They have to test each die to make sure it runs 100% satisfactorily and totally reliably when operated within specifications.
Each and every die released is tested quite intensively and graded into various clockspeed bins, they test every last minute feature of the die to ensure that there is absolutely nothing wrong with the die and it will operate wholly reliably and within it's rated specifications.

The testing procedues each die goes through is infinitely more varied and complex then any workload the processor is ever likely to see when utilized by the end-user.

Even during any given manufacturing process, the processo undergoes slight tweaks and improvements over time. These 'tweaks' generally correlate with the point in time at which stepping changes occur which also fix any known errata.

 

[AMDPwred]

Rand is right

 

[dszd0g]

CPUburn is useful if you want to see if you have adequate cooling.

 

[Mark R]

What's more is that testing done by the manufacturers' has suggested that when a CPU is used it actually degrades, and that the harder and faster and hotter that it runs, the lower its maximum speed gets.

I believe I've seen a brief synopsis of such research done by Intel (it might have been posted on this board, by an Intel engineer), and it basically says that after use, CPU maximum speed decreases (based on an average of tens of thousands of CPUs), although a very few do apparently improve.