https://bugzilla.novell.com/show_bug.cgi?id=404302 User a.vankaam@chello.nl added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c1 --- Comment #1 from Alexander van Kaam <a.vankaam@chello.nl> 2008-06-26 14:02:41 MDT --- Created an attachment (id=224668) --> (https://bugzilla.novell.com/attachment.cgi?id=224668) sensors and cpu info from 10.3 and 11.0 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=404302 User a.vankaam@chello.nl added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c2 --- Comment #2 from Alexander van Kaam <a.vankaam@chello.nl> 2008-07-02 09:49:19 MDT --- I spend this weekend doing a full clean instal of 11.0 (DVD this time not network), the temperatures still show 15c higher (which after reading up on lm_sensors) can be explained because they no longer see cpusensor in Celcius but more as a scale from 0 - 100, however what also happend was that the fans (2 back, 1 front) are also back at their 10.3 speed, about 600rpm lower then they where when I tried it with 11.0 (as shown in attachment). Left the system running for 2 days now and its seems just fine. So I think you can write this up to me boing dumb and doing something wrong somewhere along the line before. My appologies, -Alex -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=404302 User seife@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c3 Stefan Seyfried <seife@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |werner@novell.com Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #3 from Stefan Seyfried <seife@novell.com> 2008-07-02 14:03:06 MDT --- Ok. I'll still add the Maintainer of the sensors package to CC:, maybe we can improve the documentation so that this change is more prominent. - and something still was strange, because if you needed 3 installations to get your system to not get that hot, so that the fans run slower and it does not "smell hot", that's clearly a strange behaviour. However, it will probably be impossible to find out what went wrong now that it does not happen anymore. So i'll close this as "worksforme" for now. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=404302 User jdelvare@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c5 --- Comment #5 from Jean Delvare <jdelvare@novell.com> 2008-07-03 08:00:16 MDT --- Alex! I can't believe it :) It has been such a long time, how are you doing? The temperature reported by the coretemp kernel driver changed between 10.3 and 11.0 because it's fundamentally reported as a relative temperature by the CPU. As you can see, the high temperature limit changed from 85°C to 100°C because our heuristic to determine the limit based on the CPU model changed. The important thing is that the difference between the measured temperature and the high temperature limit did not change for you between 10.3 and 11.0 (85 - 36 = 49, 100 - 52 = 48.) So physically the temperature is really the same (and you have a huge safety margin.) I know it's confusing, but blame it on Intel for not clearly documenting the high temperature limit for all their CPU models. So, if something in your system was overheating, it was not visible in the output of "sensors". I have no idea why your fans were running faster, nor why they no longer do. Werner, Alex certainly has a CPU fan, just the default sensors.conf labels do not match his motherboard. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=404302 User jdelvare@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c7 --- Comment #7 from Jean Delvare <jdelvare@novell.com> 2008-07-03 09:15:53 MDT --- Internally, yes, the temperature is relative to a limit. The problem being that the limit is not clearly provided by Intel. We had to interconnect various documents and unofficial information to get it right on all CPU models. That's why we got it wrong at first and had to change it. And we still aren't 100% sure it is correct. For the user this is supposed to be transparent, as the coretemp driver itself computes back an absolute temperature in °C and exports that. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=404302 User jdelvare@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=404302#c9 --- Comment #9 from Jean Delvare <jdelvare@novell.com> 2008-07-03 10:12:53 MDT --- You can dump your w83627ehf registers using the following command: isadump 0x295 0x296 This will show you the hardware monitoring unit registers, not the Super-I/O registers. For the Super-I/O registers, either use superiotool, or isadump on the right ports (0x2e/0x2f or 0x4e/0x4f depending on your system.) If you have more questions, feel free to ask me directly. Vendor support hasn't changed a lot, with very friendly vendors and incredibly difficult ones (you know who they are, it didn't change.) What changed is that we have a few contacts in these companies now, so it's somewhat easier to get non-public information. But it can still be difficult at times. This relative temperature mess on the coretemp driver is an example of that. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.