On Thu, 08 Apr 2004, martin f krafft wrote: > I wasn't the one initiating the cross-post. Don't shoot me.
I am killing it. > also sprach Paul Johnson <[EMAIL PROTECTED]> [2004.04.08.0244 +0200]: > > That's why you should run from the battery until it runs down at least > > once a week or so > > I believe this is wrong. Lithium-Ion Batteries actually suffer from > complete discharge cycles. Any references? This is an important topic. > also sprach Henrique de Moraes Holschuh <[EMAIL PROTECTED]> [2004.04.08.0630 +0200]: > > Anything that gets the processor to "keep itself cool" while going > > from idle to almost-full-power as a crazed rabbit will fuck the > > entire system much sooner, and reduce the stability too. > > What? So cooling is bad? Cooling? If it changes the temperature too fast, maybe. But that's not anywhere a common problem. The issue here ain't cooling. Something like a 20A high frequency square wave on the onboard voltage regulator (VR) output is. That's extremely bad if the VR was not especifically designed to handle such loads. Which is expensive. For most boards, you can imagine that you won't have much leeway on what the VR is capable of handling: either it was designed with something in mind (and you paid for it), or it will perform "almost well enough" when that something happens, and cause unpredictable crashes, or even fail to perform at all. Note that boards aren't supposed to be designed for northbridge disconnects: if the manufacturer never enables that function during POST, the product "doesn't have that feature", and enabling it is a bad idea. So, if your motherboard ain't of the best designed crowd, the northbridge disconnect operation will crash it plain and simple. Look for the magic "make your athlons run even cooler" tricks for Windows of a while back *that required that you change something in the PCI config space* of the northbridge, and check the trouble and failure reports. I don't know of a similar thing for Pentiums, but that means nothing: I didn't even check, since my D875PBZ board is rock solid, and my previous ASUS A7V wasn't (which lead me to hunt a lot of information on the issue). The 2 old FIC PA2007 I had, OTOH, lasted 4 years without a glitch, and never locked up until they VR started frying SDRAM memory modules on power up. Both boards started doing it in a small timeframe, so I suspect a fab/component or design shortcoming. So yes, IMHO such types of cooling are BAD in some motherboards, and a last-resource option in good ones you want to last a long time. You need cooling? IMHO, get a proper thermal management solution, throttle down the CPU during extended periods of low activity, etc. If you never mucked around with anything in the northbridge, then you're using whatever the manufacturer thought best. If it is a well designed board, and it didn't use crappy capacitors from some industrial pirate or black market components, that won't reduce its expected lifetime any further (because it is already computed in the expected lifetime). If the board can handle it, but it is not as stable as one might want, try disabling the feature. Google search for it, and "man setpci" for more details. Laptops are supposed to be engineered for the best low-power operation that is possible, so we can leave them out of this subthread. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
