LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware (http://www.linuxquestions.org/questions/slackware-14/)
-   -   generic kernel for -currrent with 8GB ram: how much of a time saving? (http://www.linuxquestions.org/questions/slackware-14/generic-kernel-for-currrent-with-8gb-ram-how-much-of-a-time-saving-4175471086/)

textillis 07-26-2013 11:19 PM

generic kernel for -currrent with 8GB ram: how much of a time saving?
 
And how much difference would I notice in generalized speed of system:
  1. - opening/closing files;
  2. - opening/closing apps; and
  3. - using sbopkg & other package tools
(This question is sparked by an issue raised by me re. slowed startup time for X due to recent creation and editing of config files,http://www.linuxquestions.org/questi...es-4175471087/
Thanks

TracyTiger 07-27-2013 01:10 AM

Quote:

Originally Posted by textillis (Post 4997677)
And how much difference would I notice in generalized speed of system:
  1. - opening/closing files;
  2. - opening/closing apps; and
  3. - using sbopkg & other package tools

Hi textillis !

Rather than asking about performance differences, why don't you configure the different scenarios/configurations you're interested in and then tell us all what performance differences you noticed.

textillis 07-27-2013 01:25 AM

> Tracy Tiger
Good suggestion.
The very fact that I asked the question belies a lingering reticence which I have no reason to keep and every reason to jettison!

If I can configure my own conky and use my own aliases, and keep my system -current, I guess it is high time I switched to full play/experiment mode in slackware, then bring home the goodies to others.

Thanks for the gentle, timely shove

solarfields 07-27-2013 04:09 AM

i don't think you will see any noticeable difference. The machine may boot a bit faster with the generic kernel.

textillis 07-27-2013 06:15 AM

>Solarfields

That is what I thought I had gleaned from the bit of reading I'd done on the subject. But as Tracy suggests above: why not try it as an experiment?

I'll post back just in case its of interest to someone at some time.

H_TeXMeX_H 07-27-2013 06:23 AM

It will definitely boot faster. For more performance in other things you'll have to compile your own kernel. However, the difference here is also small, so you may not notice a difference.

textillis 07-27-2013 07:11 AM

>Solarfields

That is what I thought I had gleaned from the bit of reading I'd done on the subject. But as Tracy suggests above: why not try it as an experiment?

I'll post back just in case its of interest to someone at some time.

textillis 07-27-2013 07:13 AM

>H TexMex,
Thanks. I won't be putting myself through that rigorous process unnecessarily :)

solarfields 07-27-2013 07:58 AM

Quote:

But as Tracy suggests above: why not try it as an experiment?
Of course. Also, it is recommended that you switch to the generic kernel after you install the system. If you look here for example, you will find the following:

Quote:

Use one of the provided generic kernels for daily use. Do not report
bugs until/unless you have reproduced them using one of the stock
generic kernels. You will need to create an initrd in order to boot
the generic kernels - see /boot/README.initrd for instructions.
The huge kernels are primarily intended as "installer" and "emergency"
kernels in case you forget to make an initrd.
:)

TobiSGD 07-27-2013 10:16 AM

I often experiment with kernels and therefore the huge configuration is always my first choice, since I don't want to rebuild the initrd every time I have compiled a new kernel. I couldn't notice any difference in performance between huge and generic kernels. There may be a difference in boot time, but since I boot from a fast SSD this is not noticeable for me.
You should get a larger performance increase (although it shouldn't actually not be that large) with compiling a kernel optimized for your CPU than with switching to generic.

H_TeXMeX_H 07-27-2013 10:26 AM

You will get a greater performance increase going from 32-bit to 64-bit than from compiling your own custom kernel.

The huge kernel boots extremely slowly on older computers and slightly more slowly on newer computers, compared to generic. This is because it has to be loaded into RAM. As the huge kernel is huge, it takes a very long time on older computers.

textillis 07-27-2013 10:59 AM

generic kernel for current brings no noticeable performance increase
 
I followed Pat's readme and created the initrd-tree, the initrd.gz, adjusted lilo.conf settings by adding the appropriate line in the right spot, then re-ran lilo before going for reboot.

I didn't physically time both boot sequences, however, I am satisfied (by having had to reboot an inordinate number of times due to noob-klutz, that it is the same (certainly the time the splash on xfce is the same, though I expect this is incidental, since it is X starting up)

I guess my next target is Tobi's suggestion to build a machine-customized kernel (though even as i type those words, it seems like a "bridge too far" for the moment; maybe not; I'll have a look at it Tobi

thanks to all respondents! :)

PS incidentally: how can I be 100% sure that I am running the generic kernel?
I tried uname -a but that didn't provide such information

textillis 07-27-2013 11:51 AM

>TexMex
Already running SW64-current

>Tobi
Thanks; probably give it a miss for the time being if that is the case.

AlleyTrotter 07-27-2013 12:46 PM

When you do decide to build your own kernels I would suggest you review the script
Code:

/usr/src/linux/scripts/kconfig/streamline_config.pl
It will build a kernel with only the modules required by your system.
My kernel was reduced from 3095 modules to only 67 modules using this script
The script is called by using
Code:

make localmodconfig
Something to look at when you are bored.

OT had a chance to visit Australia in 1970. Greatly enjoyed my time there.
John

glorsplitz 07-27-2013 03:27 PM

@AlleyTrotter, do you know how much different streamline_config.pl is to mkinitrd_command_generator.sh?

thanks


All times are GMT -5. The time now is 03:19 AM.