> I've also asked you to undertake the second phase, which is to perform a
> binary-search set of kernel builds to pin down the timeframe in which
> this change occurred. It's also been suggested that the scheduler
> changes made by Bruce Evans may have impacted your system; you might want
> to bracket those changes to quickly eliminate them as possibilities.
In the very very very heat sensitve embedded product I'm working on, we have
a system that has thermal probes all over the place that we use for testing.
It's a Celeron 333(S370), 440ZX motherboard, and SDRAM DIMMs.
In a very very unscientific test, using software that's not available
anywhere outside my office, I can say that if anything, I see a 1 or 2
degree (C) *decrease* in CPU temperature between 3.2 and 3.4. (I do not have a
3.3 system to compare with to see if, in fact, 3.3 was very cool and 3.4 and
3.2 share the same brokenness).
I'm also using external temperature probes, not using an LM78 or anything
internal to measure the temperature. My honest guess is that something broke
in 3.4 that's making it screw with the temperature readings, and the bios
isn't setting it back right. :)
Kevin
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message