> I practice this concept myself. Besides being optimised for > speed because the kernel footprint is smaller, -- snip --
After a most interesting discussion in the swap size vs performance thread, I feel compelled to play devil's advocate once more. My understanding was that the size (footprint) of the linux kernel with drivers compiled in, out or as modules, didn't really have much bearing on performance once all the necessary pieces are in memory (including modules), and that having unnecessary bits loaded as modules or compiled in would make little or no difference. I accept that a larger kernel would be slower to start up with additional IO and housekeeping required. I expect the answer will be similar to the swap discussion in some ways, that is that there is a performance hit, but it's insignificant taking all other factors into account. In short, my question is, how is a lean kernel compilation going to provide an optimisation in speed? Happy to be enlightened, :) - Rog -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
