CPU is hot, the life is hard, i've no money... please attach logs and possible sources for this issue, we have no cristall balls near.

- edited -

I did some tests.
So, it is like i said last time: acpi-cpufreq isn't loaded at the startup, so CPU go at maximum speed.
After I loaded that module, CPU went always at maximum speed but, with
cpufreq-set -g powersave
it seems that CPU temperature go down quickly.
After this, I used also
cpufreq-set -f 800000 -c 0
cpufreq-set -f 800000 -c 1
(one command for each processor)
to set the current cpu speed at the minimum said by "cpufreq-info" command,
and temperature went down enough, about 44/45°C, my "normal" temperature.
How can we have the same result but automatically?
In laptops expecially high temperatures are easy to note.
It seems, anyway, that other distro have automatically cpu scaling when Linux doesn't require high computation power.

I hope that this time I'm been useful ;)

This is not a bug: http://www.chakra-project.org/wiki/index.php/Cpufrequtils

chakra-hardware-detection should detect if you run a laptop and set the proper settings:

https://gitorious.org/chakra-packages/platform/blobs/testing/chakra-hardware-detection/opt/chakra/hooks/hwdetect_4_power

Then tribe should pick it up and add it to rc.conf (line 175):

https://gitorious.org/chakra/tribe/blobs/master/scripts/postinstall-functions/job-setup-hardware

This code is old and maybe need to be reworked ...

I think I found it:

ls /lib/modules/2.6.38-CHAKRA/kernel/arch/x86/kernel/cpu/cpufreq
acpi-cpufreq.ko.gz gx-suspmod.ko.gz mperf.ko.gz powernow-k6.ko.gz speedstep-ich.ko.gz
cpufreq-nforce2.ko.gz longhaul.ko.gz p4-clockmod.ko.gz powernow-k7.ko.gz speedstep-lib.ko.gz
e_powersaver.ko.gz longrun.ko.gz pcc-cpufreq.ko.gz powernow-k8.ko.gz speedstep-smi.ko.gz

I was only searching for *.ko instead of *.ko* - should be fixed in a new chakra-hardware-detection.

Should be fixed now : https://gitorious.org/chakra-packages/platform/commit/33c999ca7a6365057b812ca26888253b7e239621

Here is a fix to use for postinstall: http://www.chakra-project.org/scripts/cpu-powerscaling-fix.sh

Read also this: http://www.chakra-project.org/news/index.php?/archives/194-Workaround-for-recent-power-problems.html

Good work Phill I red that yesterday, very very good ;)

I used this fix: "cpu-powerscaling-fix.sh". one of my cpu cores scale down and the other one remain fully active. Why? This should be impossible because both cores in my core 2 due cpu scale up and down together.

# watch grep \"cpu MHz\" /proc/cpuinfo
Every 2.0s: grep "cpu MHz" /proc/cpuinfo Mon Jul 4 01:45:54 2011

cpu MHz : 1000.000
cpu MHz : 2000.000

Seems I'm still on fixing this bug in tribe: http://gitorious.org/chakra/tribe/commit/9ba565c467be0d1914bc1a13cd994e37cbd50dad

This bug seems fixed with the last kernel.

Please report back if this issue is fixed.

This issue is again a bug

Is this nepomuk related? Still high CPU temp with nepomuk/strigi disabled?

Some comments: the code is old.

1. Tribe adds to the MODULES line of rc.conf every known module for the system, as if the kernel couldn't load them by itself.
2. Despite the fact Tribe does this, it doesn't add the proper power throttling modules to rc.conf, and it doesn't configure the system to use ACPI. Tribe is hard on you, sucks power, and you need to customize rc.conf to get good results. These are my rc.conf MODULES lines, for posterity.

#MODULES=( nvidia battery button processor thermal video btusb cdrom edac_core edac_mce_amd hid k8temp i2c-nforce2 i2c-core evdev joydev psmouse serio_raw sparse-keymap media uvcvideo v4l2-compat-ioctl32 videodev mmc_core hp-wmi wmi ssb nvidia nv_tco crc16 snd-hwdep snd-page-alloc snd-pcm snd-timer snd snd-hda-codec snd-hda-intel soundcore pata_amd scsi_mod sata_nv forcedeth bluetooth rfkill pcmcia pcmcia_core usb-storage usbhid usbcore ehci-hcd ohci-hcd uas sd_mod sr_mod st sg ac long_isnt_it? )
MODULES=( powernow-k8 nvidia )

Of course, the long_isnt_it? module is a bogus one, and the thing doesn't give an error since I commented the line :D. My MODULES line has only two modules: nvidia (I need to load it manually) and powernow-k8 (my power throttling driver, for Intel, this is acpi-cpufreq)

Also, keep in mind that appset-qt is tremendously inefficient while it's idling. In my system it idles at 2% CPU, when it should be 0. That includes periodical updates for the tray icon when it doesn't do anything, and an unbearable, CPU-eating, screen flash when appset-qt is compiling something from CCR.

@Ernesto, mind that AppSet is not developed by Chakra, and this is the right place to complain about its bugs. Please, report them here: https://sourceforge.net/tracker/?group_id=376825

I confirm this bug. Just performed a brand new installation and the acpi-cpufreq module isn't loaded. I noticed that the chakra-hardware-detection project wasn't installed also.

i can confirm this too. just fresh install on my asus x42jy laptop from chakra 2011.11 edn livecd, upgraded to kde 4.7.4 from default stable repo and acpi-cpufreq module didnt load by default. "cpufreq-info" in terminal just gave error about cannot find module for power management. solve by modprobe the module manually and editing rc.conf to include the module upon boot.

I'm also confirming this bug. HP nx7300 laptop. Also had add manually acpi-cpufreq to rc.conf.

Still nothing? This make impossible use Chakra on a laptop...both for overheating and for battery time.

I used to think it was a Core Due Duo issue but I see that also other cpu are affected.

I think we must narrow down this bug, because "CPU is hot" is a poor description of what's happening. The issue is:

- Chakra doesn't load the appropiate CPU throttling modules (powernow-k8, acpi-cpufreq) automatically. This is something it shares with Arch Linux.
- Chakra also doesn't enable the relevant energy saving facilities automatically. They must be enabled manually in all cases.

@Nicola: what you state is partially true and partially false. This bug makes impossible to use Chakra OUT OF THE BOX on a laptop. If you tweak the laptop, you can get comparable battery life with other Linux distros.

@Ernesto: there is some tutorial to make chakra working on a laptop and have good battery life?
The problem is all in the CPU scaling, you're right...but this is a important issue I think, especially for new users

There is a guide in our wiki how to enable cpufrequtils:

http://chakra-linux.org/wiki/index.php/Cpufrequtils

We work hard on Tribe atm to fix known issues. To follow the kernel-development check here:

http://chakra-project.org/bbs/viewtopic.php?id=6534

Adding insult to injury, with a recent Archimedes install the powernow-k8 and acpi-cpufreq modules are correctly detected, but they appear as "powernow_k8" and "acpi_cpufreq" in the MODULES line, and fail to load.

Cpupower is added to the testing repo's, and will eventually replace cpu-frequtils. This, combined with all the recent changes to the kernels, can anyone report to the status of this bug in combination with cpupower?

Just installed a fresh Chakra 2012.7 in my Pavilion tx1000, and there's no heat (that's great, since if I suffered from this bug, my tx1000 would be dead by now). All the relevant powernow modules are loaded, so I'd consider this as fixed with Chakra 2012.7.

Closing this bug now, anyone still having an issue, please start a new report, after you switched to cpupower.
Most of this report dealt with cpu-frequtils, package is no longer maintained, and will be depreciated in Chakra.