Linux - NewbieThis Linux forum is for members that are new to Linux.
Just starting out and have a question?
If it is not in the man pages or the how-to's this is the place!
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Long version- you've got a brand new release laptop, using a GPU intergrated into the CPU system (which AMD is calling an APU). A lot of the kernel code to make everything work properly has only starting going into kernel 2.6.38, that is ongoing with kernel 2.6.39. Driver work from the open source and AMD catalyst closed drivers is also ongoing.
Yes, but there is a complicating factor. That is, there have been kernel power usage regressions recently (2.6.38 particularly, but also an earlier one on 2.6.35). There is a suggested work-around for this, but only some systems will be affected...and maybe even the work around won't work on all systems, so the picture there is a bit unclear. A lot of the measurement work has been done at Phoronix and there are articles describing the situation here and here:
It even sounds as if the first of the 3.0 kernels won't have the fix in officially, although probably some distros will be able to put in the external work-around in 'laptop mode', or whatever.
So, I was intrigued by this thread, and decided to do some homework, and a little testing.
I downloaded powertop to check my power consumption while running a handful of distros. I ran all of this on my Dell laptop, which has a Core 2 DUO 2.20Ghz processor, 2GB RAM, nVidia Quadro NVS 140M board (I was not using the nvidia proprietary drivers for any of these tests), and an Intel ICH8 chipset. My laptop has a single harddrive that is fully encrypted and used by windows, so it was not accessible to any of these distros; they were all run off of my 8GB Kingston data traveler 101 G2 flash drive, with NO SWAP partition in use.
The version of nouveau being used will imapct on the power consumption. IIRC the newer versions have better support and are less likely to 'waste' watts.
Quote:
Originally Posted by Ahau
What my test seems to indicate is that, for my system at least, the size of the distro, the RAM usage, and the desktop environment are relatively arbitrary. There is only a 10% spread between the highest and lowest consumers, and the lowest consumer was the largest, most RAM intensive distro.
Interesting result. I'd love to see it backed up with more hardware power consumption stats (like jefro's suggestion of a wattmeter, though that can skew results).
Quote:
Originally Posted by Ahau
NOTE: I also tried lowering my CPU to 800mhz, in Porteus. That didn't seem to change the power consumption. I'm looking into that, to see if it's related to my processor and/or my BIOS settings.
Forcing the CPU to limit itself to 800MHz wont make a huge amount of difference at idle, the CPU should already be throttled down. If you went into the BIOS and dropped the CPU voltage as well, you should start seeing a bit more power savings, but it wouldnt really be that workable without limiting the CPU to 800MHz full time, and/or doing a lot of 'overclocking' systle tests to see how much voltage you can remove and still have the system run stable.
Quote:
Originally Posted by Ahau
I just ran this on the distro's I had already running around, so my apologies for not doing a more exhaustive, broad test.
Any data is good. Sure, I'd like to see more testing done, but thanks a lot for the data you have already collected. BTW, I'd consider this to be worth a blog post or its own thread.
Quote:
Originally Posted by salasi
Yes, but there is a complicating factor. That is, there have been kernel power usage regressions recently (2.6.38 particularly, but also an earlier one on 2.6.35). There is a suggested work-around for this, but only some systems will be affected...and maybe even the work around won't work on all systems, so the picture there is a bit unclear. A lot of the measurement work has been done at Phoronix and there are articles describing the situation here and here:
It even sounds as if the first of the 3.0 kernels won't have the fix in officially, although probably some distros will be able to put in the external work-around in 'laptop mode', or whatever.
True. I just thought that it was worth mentioning, especially as AMD fusion wont be able to use earlier kernels (well, not for APU support anyway). Hopefully the developers get these power regesssions out, sooner rather than later.
I don't have a wattmeter, and my goal is to reduce power consumption while I'm unplugged. I suppose I could buy a wattmeter and change my power profiles so that my "plugged in" settings match my "unplugged" settings (and I might, if I get really deep into this). Thanks for the tip!
I recompiled my kernel for Porteus yesterday and tried adding some of the suggestions that I've found on the web. However, I think there have been enough recent changes in the kernel (I'm using 2.38.8) that powertop needs an update before it will really be able to give me more data (it tells me that it needs CONFIG_TIMER_STATS enabled, and I can't find that in my config menu or .config file) . So, the kernel recompile (and compiling powertop against it) didn't really do me much good.
In other news, I downloaded WattOS and burned it to a CD. That had about the same power consumption (again, according to powertop) as the distros mentioned above, but at least powertop was giving me some more information.
In all cases, I've found that my LCD backlight is a power hog. When turned all the way up, it uses nearly 10 watts (vs. having it turned all the way down). Enabling USB autosuspend can save me another 3 watts or so. I'm going to try to figure out how to do some of these things outside of powertop so that I can put them in a startup script. Hopefully that program will get upgraded for more recent kernels.
@cascade9, thank you for your comments!
EDIT: I think I've found CONFIG_TIMER_STATS...stupid me was looking for that in my config file, when powertop specifically says it's CONFIG_DEBUG_KERNEL in the config file... more to come, once I have time to recompile.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.
