LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - General (http://www.linuxquestions.org/questions/linux-general-1/)
-   -   Linux kernel version with best battery life? (http://www.linuxquestions.org/questions/linux-general-1/linux-kernel-version-with-best-battery-life-919879/)

mahkoe 12-21-2011 03:01 PM

Linux kernel version with best battery life?
 
I'm compiling and installing the latest linux version (3.1.5) on my 64 bit Ubuntu 10.10 install no matter what, but I've been reading several articles that say that the newer kernels have crappy battery life, but there is this magical patch that fixes everything, and on top of that another magical unrelated patch (the ever-so-famous 200 line one) that lowers the latency.

I was wondering if someone could set the record straight and let me know which version has which patch and which kernel version is the best for battery life.

Thanks

salasi 12-21-2011 03:43 PM

Quote:

Originally Posted by mahkoe (Post 4555588)
I was wondering if someone could set the record straight and let me know which version has which patch and which kernel version is the best for battery life.

Here is what I know, or I think I know:
  • The famous '200 line patch' has, as far as I know nothing to do with the 'battery life problem'; the 200 line patch is only really useful in one use case, and that happens to be quite similar to the use case experienced by a lot of developers. Coincidentally, or otherwise, this patch seemed to get something of a fast track.
  • Phoronix has been collecting some numbers; if you want numbers go to Phoronix.
  • There is argument that the Phoronix interpretation, or more exactly, their use of language is wrong/wildly deceptive/unfortunate. I'm with that argument, and it is not just a mere semantic argument, as what you think of that argument might change what you do.
  • <opinion> It isn't a regression, at least not a kernel regression. With the previous approach, there was some hardware that just didn't work correctly. The fix for the misbehaving hardware, increased current consumption. There was then a fix for the fix, that, effectively only applied the first fix in appropriate cases, which was a bit tricky as part of the original problem was hardware not reporting its capabilities correctly, and thus deceiving the software about in which cases the fix should be applied. While this is a bit of a mess, really it was almost inevitable, given what happened with the hardware </opinion>
  • I am a bit unsure about this, but I believe if you do see the increased consumption on the current kernel, it is because you are using hardware that is buggy; with non-buggy hardware, there isn't a (significant) climb in current with the current kernel. I don't know exactly what goes wrong if you use an older kernel, but you could find weird things happening.

mahkoe 12-21-2011 03:56 PM

Okay. So, what you are telling me is, the 200 magical patch may not be all it's cracked up to be, the 3.1.5 version has updates that are good for battery life, and that this excess power consumption is due to buggy hardware.

So, let's pretend I have my current kernel (2.6.35-31) and my self built one (3.1.5) and I don't have any way to charge my laptop, and I'm in a life and death situation where I need more battery life. Which is better, and would some of the older kernels give me a better chance of survival, if so, which?

salasi 12-21-2011 05:15 PM

Quote:

Originally Posted by mahkoe (Post 4555645)
the 200 magical patch may not be all it's cracked up to be

Well, you are free to believe whatever you want to believe, but my formulation would be closer to 'it won't be useful for most people, but it will be well worthwhile for the minority for whom it is an advantage'.

Quote:

Originally Posted by mahkoe (Post 4555645)
So, let's pretend I have my current kernel (2.6.35-31) and my self built one (3.1.5) and I don't have any way to charge my laptop, and I'm in a life and death situation where I need more battery life. Which is better, and would some of the older kernels give me a better chance of survival, if so, which?

Well, obviously you should measure it, under your use case and with your hardware. After all, it is your life or death.

If the bookies were offering odds, I'd bet on your older kernel by the smallest of margins, but buggy hardware could turn that into a wide margin.

mahkoe 12-21-2011 05:45 PM

Okay then. Well thanks a lot, I'm marking this thread as solved


All times are GMT -5. The time now is 04:35 AM.