View Single Post
Old 09-03-02, 04:41 PM   #5
HotCarl
Registered User
 
Join Date: Aug 2002
Posts: 13
Default

First things first. All the software in the world is not going to give you accurate numbers since the thermistors(temperature probes) built into the chips all have 'error' in them. The only accurate way to get a reading is to use some form of external measuring probe. That's why the folks at Anandtech in the link above used an InfraRed temperature probe.

Now, when is hot = too hot? When the GPU gets too hot it will cause various geometry in a game to 'pop' into incorrect shapes and colors. Lockups are not too uncommon either. When video RAM gets too hot it causes colorful, sparkling pixels all over your screen. All this heat is only going to be generated when the card is taxed so don't expect problems to occur at the desktop.

As your card is running in AGP 4x mode, I will assume that the AGP port is 2.0 compliant. No problems there. You will always observe fluctuations in the voltage supplies as the loading on the main power supply changes. Yours changing from 1.54 to 1.56 and 3.29 to 3.31 is well within the norm. If you have a power supply that is insufficient for the load produced by all your devices(70% of the Supply's max output is the safe, constant load for it), it will overheat and provide less than acceptable voltages and stability.

Get good airflow thru your computer's case. You don't feel cool on a hot day and neither does it!

As for overclocking, try RivaTuner or any of the GeForce type tuners. Check www.guru3d.com for a few popular ones. Oh yeah, overclocking consumes more current and produces more heat.

Enjoy!
HotCarl is offline   Reply With Quote