Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

Tweaking the kernel

New to Debian (Or Linux in general)? Ask your questions here!
Post Reply
Message
Author
JDVW
Posts: 6
Joined: 2017-06-26 16:42

Tweaking the kernel

#1 Post by JDVW »

Does anyone here modify their kernels for better performance? I read somewhere today that it can be done. Something like removing unnecessary drivers.

User avatar
stevepusser
Posts: 12930
Joined: 2009-10-06 05:53
Has thanked: 41 times
Been thanked: 71 times

Re: Tweaking the kernel

#2 Post by stevepusser »

A driver won't be loaded if the hardware is not detected, so.... You can speed up compilation and make the kernel package smaller by pruning unused drivers, but that's about it on that front.

You could get into tweaking schedulers and latency like the Liquorix kernel, though.
MX Linux packager and developer

JDVW
Posts: 6
Joined: 2017-06-26 16:42

Re: Tweaking the kernel

#3 Post by JDVW »

So the Liquorix kernel looks like a Hybrid kernel.

Bulkley
Posts: 6383
Joined: 2006-02-11 18:35
Has thanked: 2 times
Been thanked: 39 times

Re: Tweaking the kernel

#4 Post by Bulkley »

Many years ago tweaking a kernel could produce noticeable results. That was back when a "blazing fast" CPU was a 486.* Those days are gone. On modern machines you can tweak to your heart's content but to see the results you need before and after bench tests.

* I'm remembering a magazine cover that boasted "blazing fast" 486.

User avatar
dasein
Posts: 7680
Joined: 2011-03-04 01:06
Location: Terra Incantationum

Re: Tweaking the kernel

#5 Post by dasein »

stevepusser wrote:You could get into tweaking schedulers and latency like the Liquorix kernel, though.
+1

There's very little that Liquorix hasn't done to try to optimize performance for desktop (as opposed to server) use.

That said, I echo the general sentiment that hardware is "the" driving force in kernel performance, and has been for at least 10 years.

I've not done any defensible research on the question, but my educated speculation is that Liquorix's "performance boost" is largely perceptual; that is to say, it appears faster, subjectively speaking, but in reality, the total time that a given takes to execute is still roughly the same, start-to-finish.

But again, that's just speculation. (And just to be clear, apparent speed-up is not a bad thing.)

Post Reply