• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Electromigration - Interesting Read

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

LarryJoe

Member
Joined
Feb 19, 2002
I was posting in a thread at Anands from a member who was concerned about going to 1.85 volts on his 1.6A. Of course I got into.....um.....let's just say the typical difference of opinion with one of the common AT know-it-all'ers.

KenAF posted in the thread in resposne to the quote below and wrote this very well written explanation of Electromigration:

<< Read the articles on it...Too high of voltages can cause a term called electromigration and this will kill the chip...Quit telling ppl "if the temps are fine", cause you don't seem to understand the whole picture... >>

If I hear this one more time, I am going to go nuts. This all arose out of a single comment, or more appropriately, taken out context by Johan at Aceshardware.

Do you people even know what electromigration is? Electromigration is normal; if it weren't for electromigration, processors might last tens or hundreds of years. The danger of electromigration is that the processor will eventually require more voltage to remain at the same speed, and will need to be reduced to a slower speed at the given voltage to remain stable. Eventually, after a substantial period of time, it may stop working completely. However, this isn't something that could occur to a processor overnight, or even over a period of many months. Even at 1.85V on a P4 (assuming reasonable temps), it will probably take years.

Electromigration occurs just from running processors at 100% usage. It occurs when running processor intensive tasks. Some have suggested that electromigration occurs (on the Athlon, not just the P4) whenever the processor temperature exceeds 50C.

Primary electromigration risk occurs from increased temperatures; failure from electromigration increases exponentially with increased temperatures. That said, the copper interconnects in the P4 can carry ten times as much current as aluminum wire, and serves to allow substantially increased temperatures and electromigration without risk to the processor. That's part of the reason why Intel moved from aluminum to copper interconnects with the Northwood; so they could reach higher clocks, and tolerate more heat, while still preserving long-term (essentially indefinite) reliability.

What's a real world example of electromigration? Well, people who overclocked their Celeron 300A @ 581MHz have begun to find, after years and years, that their processor no longer runs stable at that speed. Some of those who ran their Celeron 300A @ 581MHz for several years now find that it only runs stable at 450MHz. That's due to electromigration. Actual processor failures due to electromigration are very very very rare, considering that very few people keep their processors for more than five years.

What is the worst danger one can realistically expect to face from electromigration with the "Northwood" P4? Well, if you don't have good cooling, you may find in three years that your P4 1.6A will no longer run stable at 2560MHz. Instead, you may have to clock it down to 2400MHz. Big deal. The chance of a P4 processor, with increased voltage and adequate cooling, completely failing in several years is about as close to 0% as you can get.

--------------------------------------------------------
I thought you folks might find this interesting. Here is a

link to the thread.
 
Last edited: