Basic overclocking; relationship between voltages and temps?

[nicky9499]

Hi guys,

I am new to overclocking, turning to this method as a stopgap measure before a new build perhaps this year end. The hardware in question is old stuff, an Asus M2N68-VM and AMD Phenom1 X3 8450 2.1GHz with stock HSF and Thermaltake paste. Currently at 2.410Ghz (229.57 x 10.5) and tops out at 50*C on full load. Testing with Prime95 and monitoring with Core Temp.

What I'd like to ask is would it be preferable to have a higher multiplier and lower base speed or vice versa? Haven't touched voltages yet (stable so far) but I am only able to bump in 50 millivolt increments. How much is this "jump" in relation to the base speed and how will it affect temperatures?

On a sidenote, Windows Explorer still reports me as having 2.1GHz, what's up with that?

Cheerio,
Nicholas

 

[Big B]

It's not an exact science, so what one person ends up with may not even be close to what you get.

Moving up with just the multiplier offers the advantage of only messing with the CPU. The bad news is that the jumps in clockspeed are large---not good for fine-tuning---and some CPU's multiplier are locked.
The FSB (or whatever terminology is used per platform) is good as well. It's based on the multiplier, so if you're determined to get that last MHz out, this offers a closer means to do so. In the past, this was locked to a ratio so you could keep PCI devices happy, but I'm pretty sure chipsets today don't have that predicament.

Don't increase the voltage unless you need stability. See how far you can go before hitting instability, then bump up the voltage in as small of increments as possible. As the clockspeed increases so does the temperature. Additional voltage puts extra heat in the equation.

As far as Windows Explorer not showing the correct CPU speed, make sure you're saving before exiting BIOS and check the speed with an additional program like CPU-Z.