Well, Thanks for your good job.
But I have some thing below:
1) For ubuntu 10.04, powertop 1.96 doesn't seem work right.
Package | CPU 0
Idle 100.0% | Idle 2025.1%
1203 Mhz 0.0% | 1203 Mhz 0.1%
2.41 Ghz 0.0% | 2.41 Ghz 0.0%
| CPU 1
| Idle 2025.2%
|
|
my platform supports 4 frequencies, it shows only two. And 2025.2% is a
strange number.
Powertop 1.13 is ok.
2)
Another problem, what's the sampling interval for powertop 1.96 ?
For 1.13, if Itype on the keyboard quickly, I can see the P state frequency
changes quickly.(because I use ondemand governor)
But 1.96 is static. I don't think it is right.
At 2011-01-03 03:50:50，"Arjan van de Ven" <[email protected]> wrote:
>Hi,
>
>I just cut the second alpha of the 2.x series of PowerTop:
>
>http://www.kernel.org/pub/linux/status/powertop/powertop-1.96.tar.gz
>
>The biggest changes are around building and installing on various distros,
>better translation support, and one new feature: PowerTOP tries to show which
>process is playing audio, and will try to "charge" the power the audio codec
>takes
>to this process.
>
>I'd like to thank those who gave feedback on the first alpha release, it is
>very much appreciated,
>and I hope you can see the result of your feedback in this second alpha
>release.
>
>I've included a git shortlog below for those who want to see the nitty gritty
>details ;-)
>And as always, you can get the git tree from
>
>git clone git://git.kernel.org/pub/scm/status/powertop/powertop.git
>
>- Arjan van de Ven
>
>
>
>
>Alexander Monakov (1):
> Fix Makefile mistakes
>
>Arjan van de Ven (22):
> mention the iw code in the README
> don't crash on a NaN
> include the nl80211.h kernel header because Debian ships an ancient one
> that isn't good enough.
> use proper UTF-8 for printing Watts as well
> fix goof in makefile
> One more NULL pointer crash fixed, curtosy of Michael Riepe
> minor cleanups in preparation of showing a per tab bottom line
> show that<Enter> is a recognized key in the Tunables page
> silence a not-useful error message
> Lots of changes to support parsing who owns /dev files, and charge the
> power of the device to the process...
> fix some valgrind warnings by just simplifying the code a ton
> fix a bluetooth tunable crash
> show who has the audio device open
> write a list of processes that keep devices open to the html output
> try to estimate how many minutes of battery are left
> fix two file descriptor leaks
> more translatable strings
>
>Sergey Senozhatsky (5):
> reset terminal state (resend)
> cpu/* non-throwing new operator
> initialize perf name pointer in contructor
> reset_display on intel get_msr error
> cpu/*: non-throwing operator new (2)
>
>
>_______________________________________________
>Discuss mailing list
>[email protected]
>http://lists.lesswatts.org/listinfo/discuss
_______________________________________________
Discuss mailing list
[email protected]
http://lists.lesswatts.org/listinfo/discuss
