On Sun January 13 2008 6:37:32 pm you wrote: > The coretemp readings come from a digital temperature sensor inside > the processor. The processor provides PCI device entries which access > these internal sensors. > > The thermal_zone readings come from ACPI, so they might come from any > source, but 99% of the time, they are analog temperature sensors on > the processor, read by converting from analog to digital on the > motherboard. The processor does have a temperature sensor pin for each > core, which is this analog sensor. One would think therefore that the coretemp values would generally read higher than the ACPI values, given that the coretemp sensors are embedded in the core? This isn't the case with mine however - the thermal_zone values are always significantly higher. <SNIP> > What does this mean for you? Well, you don't want any of the sensors > to go too high, since they are all measuring real temperatures. Which is another area of confusion to me. I have tried to find out what the recommended temp range is for my processor, but when I emailed Intel they told me they couldn't help me as my processor was an OEM product not a retail product. They suggested I contact Acer, which I did, but never got a response, so I'm still flying blind as far as knowing what temps would be considered "high" or " cause for concern" for my processor. > Also, you might be able to cool things down faster by watching the > coretemp sensors, since they respond more quickly to changes in the > load on the chip. But your CPU fan is probably watching the analog > sensors, so it will be slow to react to a sudden change in CPU > temperature. This is interesting. What puzzles me is that the thermal_zone values are nearly always exactly the same for both cores, whereas the coretemp values are almost always slightly different (which makes more sense to me) > But don't feel like high temperatures mean instant death. Intel and > Acer have probably taken good care to be sure the chip can't be > permanently damaged -- in fact, if the chip actually gets close to > damaging itself, the laptop will probably turn off, instantly, without > warning. Well that's good to know. Many thanks for the excellent explanation David. Much appreciated :-) ~ Carren
