Going to one HD 4546

Jean-David Beyer

Put bluntly, you don't know what you're talking about. There are reasons why the overwhelming majority of heatsink fans are designed with fans blowing downward onto the heatsink fins. For one thing, they actually do indeed work, amazingly enough.

The stock heatsink on even the fastest hottest Intel P4 Prescott core is of the same traditional heatsink-fan design as always. The fan blows downward onto the heatsink fins. The Prescotts generate more waste heat than just about any processer ever designed.

Wrong. Intel has stock Xeon heatsink-fans with the traditional layout:

These are not suitable for blade servers because of the physical constraints of a blade, but they're perfectly fine for a tower.

Maybe so, maybe not. You don't know it.

He also perceived the wattage rating on his PSU to be a problem--but he was wrong. Many of his other changes seem to be motivated more by capriciousness and a desire to tinker than anything else.

I get the impression that he has a vague idea that "cooler is better", but he doesn't have a firm idea of WHY cooler is supposed to be better.

Nonsense. If the MTTF is already far longer than the expected useful lifetime of the equipment, then who cares? I wouldn't expect to be using a particular CPU for more than a decade or so. And if a CPU fails after a few years, so what? By then replacing it is dirt cheap.

It happens all the time with PSUs in traditional layout computers. It's a serious problem, because CPUs generate more heat than PSUs, and PSUs require lower operating temperatures than CPUs for reliable operation.

PSU failures are also more problematic than CPU failures. The symptoms of PSU problems are more difficult to recognize and diagnose, and they often cascade into causing mysterious hardware failures in other components. Worst of all, PSUs have greater potential for longer useful lifetimes, and don't depreciate in value quickly the way CPUs do. Replacing an old PSU costs more than replacing an old CPU.

You continue to repeat this straw man. Quiet PC enthusiasts don't skimp on reliability.

So what you're saying is that you have ~14 degrees to play with and you simply chose to not run it anywhere near the limits. That's your equipment and your decision. My equipment is different, and my decision is different.

Me? My main workstation is a 2.5ghz Northwood. For various design reasons, the CPU ended up being at the end of the airflow chain, receiving the warmest air. The only fan in the system is an 80mm fan running silently at 5volts. I have a switch to flip it to 12v for reencoding video files, but other than that I run at 5volts. The machine is perfectly stable at all times.

When I upgrade to 64bit, I'll be using an even cooler processor (probably a Venice core A64 unless something even better comes out). At the same time, it'll be using even better heatsinks that weren't available when I put together my Northwood, and the PSU will be significantly more efficient.

And yet you dogmatically advise us to blindly obey Intel's recommendations also.

Since that's your equipment and your decision, that's fine for you. But other people are other people, with other equipment. I'm sure the OP's processor never cost him $750, ever, and certainly none of mine ever cost that much.

Isaac Kuo