AMD Optimal BIOS settings + Overclocking Guide

JMke writes "Here's a step by step guide on how to get the most out of your AMD setup. Overclocking tips and BIOS tweak settings discussed, as well as an overview of the more popular overclocking tools. Start your overclocking here!" Lots of good info here for getting the last bit of performance out of your system while also watching out for dangers that could fry your processor.

Comments

Temperature

[DrEldarion]

From the article: NEVER LET IT GET ABOVE 55C IN MY OPINION ON AIRCOOLING

When I had the stock heatsink/fan on my 1800+, it would IDLE at over 55C, and peak at around 63C under a full load. I never had any problems. Are overclocked CPUs more sensitive to heat, or is this just a "to be safe" recommendation? AMD says they're good up to 85C.

Overclocking

[Ianoo]

The biggest thing I've always found with overclocking (both Intel & AMD) is that 'YMMV' (your mileage may vary). I've talked to people who say they run their systems at 20% more speed than they should be doing, yet one of my systems constantly locks up even if midly overclocked. Increasingly I wonder, given the current prices for this kind of technology, whether overclocking is a bit pointless. Do I spend $100 extra or do I spend 12 hours mucking with my BIOS? You decide!

Overclocking also seems so 1990s now, most people I speak to who rate themselves equal to my own geekdom are in to case modding and quiet PCs... concepts which I find much more interesting than getting 5% extra FPS in Quake.

Re:Overclocking

[gl4ss]

the current pricing of cpus however would indicate that your hypothetic 2.1ghz would cost 800$.

AMD Duron 1600 MHz, 64 KB 50.7 e
AMD Athlon XP B 2800+ 2.08G/333 MHz, 512 KB 156.0 e

(some of the disabled cache is re-enableabe on those durons, afaik, and seemingly most do run 2.2ghz+ or even 2.4ghz)

Intel P4 HT 2600 MHz s478 Northwood, 512 KB, 800
boxed 198.2 e

Intel P4 HT 3200 MHz s478 Northwood, 512 KB, 800
boxed 472.5 e

for most though the point is that they can't afford that high end cpu, and for the rest is that they don't want to pay such prices you end up buying for high end cpu(that can't be overclocked in the same ratio as cheapos can). the cooling costs anyways, regardless of if you're going to use it for oc or not(unless you're willing to go with a cheapo, or one that comes with boxed cpus).

hell.. i don't make my buying decisions based on overclockability, i make them on other factors. i just oc them as far as they're willing to go after i've bought them(well, i got one underclocked machine too, as nat-box).

Re:Overclocking

[imnoteddy]

'YMMV' (your mileage may vary)

This is true. The way to understand the YMMV factor has to start by understanding how integrated circuits (chips) are made. They are produced on wafers - flat, round thin slabs of (usually) silcon - typically 200 millimeters (mm) or in some cases 300 mm in diameter. The wires and transistors are created during dozens of steps that involve exposure to light, vapor deposition, chemical etching, and other processes. Many steps involve heating and cooling the wafer.

So why is there variation among chips? Because the process cannot be controlled perfectly, though the manufacturers try really hard. For example, chips on the edge of the wafer will cool off faster than chips in the middle. There are so many variables to control that some will be off. Twenty percent variation is very common, and designers often do simulations based on that amount of variation.

People design chips to a given clock rate, then test to see what they get. A simple example is the Power PC (PPC) that IBM makes and Apple markets as the G5. Apple sells systems that run at 1.6, 1.8, amd 2.0 GHz, but all PPCs come from the same design. The chip in a 1.6 GHz G5 may run at 1.79999999 GHz if given the chance, but it also might fail at 1.61 GHz. Some chips might test out at 2.2 GHz, but probably not enough to sell.

Re:Fry?

[Scottaroo]

Tom's Harware loves doing stunts like this. They've run processors without coolers just to see what happens, and they investigated problems with the early AMD thermal diodes. Check it out