Slowly overclocking

[ChrisC]

Ive heard people say that when overclocking, it is best to overclock a bit, then, let it sit, before overclocking higher. Their argument was that it allowed the CPU to get used to higher speeds, and going slowly would aid in final overclocability, and help the CPU to last longer.
Does anyone know if this is true, or if it just started to keep people from overclocking till their CPU melts?

[usasma]

Dunno if it's true or not. Most arguements are, as you said, that the CPU needs time to settle in. I read one thing about it just being a chance for the thermal paste to settle in - different heat amounts=different amounts of expansion.

I do it slowly to make sure that my system isn't screwing up in minor ways - I'd like a little advance warning before my system locks up. Previously I've had to leave my case open so I could short the CMOS as I backed off of the manic o'clock attempts.

2.53 gHz P4 now at 2.75 gHz and holding (145 mHz FSB) w/o problems for 3 days

[exeter_acres]

From an engineering standpoint...you should be able to just go for it....

However, my experienc si that a good long burn in period helps greatly...
by burning in, I was able to achieve higher overclocks with less core voltage than without a burn in...

I usually will run Prime95 for AT LEAST several hours...and then I run Folding@ Home 24/7... games, etc. etc. and don't try to push the next FSB jump for at least a couple days.....

Works for me
There are people who will tell you that this is bunk... I am just passing along what has worked for me and others....

[dRiNkEr]

short the CMOS ?? what is that?

[fleppen]

resetting the BIOS by changing a switch on the mainboard or by removing the BIOS battery for a couple of minutes.

[usasma]

I just remembered reading an article in MaximumPC mag that claimed that increasing your core voltage for a while, before overclocking, will also help stabilize your system for overclocking. They used some of the same reasons that have been stated above.

I'm guessing that this is all related to throwing something hot into ice cold water. It may just shatter, or it may temper it and make it stronger. Whatever it is, there's enough circumstantial evidence out there that it may work...better safe than sorry!

[fleppen]

upping vcore should only be done when a chip isn't 100% stable at a particular speed, not beforehand.

[ChrisC]

Ya, upping the Vcore will make your chip run hotter, but that is only part of the story.
An article I read recently pointed out electrical interference, in that too much Vcore will cause stronger magnetic fields within the chip (magnetic fields are caused by moving electricity), which would affect electrons elsewhere in the chip. Thats why, as the production process gets smaller (.18micron to .13 micron, for example), voltages go down. This is because, if voltages stay high, electrical interference will occur in the chip, so, voltages go down. Also, higher voltages can break silicon transistors, causing them to stay open or closed. Or just melt some circuits.
So, Vcore can be dangerous. Still, it is necessary to up it sometimes, and should not pose a problem. The chip should run out of speed before it reaches really dangerous VCORE levels.