Patching kernels, moving shadows.
Friday, 5. October 2007, 15:39:13
So long and thanks for all the fish...
Too sad - http://members.optusnet.com.au/ckolivas/kernel/ - the -ck patchset has ended for good. One good effort to put more focus on Linux for desktops. An admirable effort, which impresses, if not with accurately measurable results, with user feedback.
I am quite sad, because accidentally one of the efforts of -ck was a scheduler idea that helped me a lot in my OS course. After coding nice funky things like system calls over the skeleton for a BSD-like system, I had to come up with a couple of selectable scheduling algorithms and run a couple of tests on them. For one reason or another all the implementation of OS161 included alternative, simple-yet-effective solutions. The scheduler I was looking for was exactly that. The staircase idea from -ck - was coded pretty smoothly and with no hassle of doing tons of useless complicated things. I didn't use a lot of linux on my computers at that time, so I hadn't the slightest clue about the ck patchset, until very recently, when someone linked me the article in which he explains why he dropped the whole thing. Well - I think he is very very correct, painfully correct even. And this is why I decided to include his patchset in my kernel on my new Xubuntu. And so I did.
And I have to tell you - I don't care about what the other kernel guys have to say about measurements, it's painfully clear when I use it that the responsiveness and focus is much much better.
I've the ubuntu package here - standard ubuntu compile flags, ck patchset, ready for install. If you want it and you don't want to build it yourself - I'm guessing that it should work just fine. It's worth it.
Too sad - http://members.optusnet.com.au/ckolivas/kernel/ - the -ck patchset has ended for good. One good effort to put more focus on Linux for desktops. An admirable effort, which impresses, if not with accurately measurable results, with user feedback.
I am quite sad, because accidentally one of the efforts of -ck was a scheduler idea that helped me a lot in my OS course. After coding nice funky things like system calls over the skeleton for a BSD-like system, I had to come up with a couple of selectable scheduling algorithms and run a couple of tests on them. For one reason or another all the implementation of OS161 included alternative, simple-yet-effective solutions. The scheduler I was looking for was exactly that. The staircase idea from -ck - was coded pretty smoothly and with no hassle of doing tons of useless complicated things. I didn't use a lot of linux on my computers at that time, so I hadn't the slightest clue about the ck patchset, until very recently, when someone linked me the article in which he explains why he dropped the whole thing. Well - I think he is very very correct, painfully correct even. And this is why I decided to include his patchset in my kernel on my new Xubuntu. And so I did.
And I have to tell you - I don't care about what the other kernel guys have to say about measurements, it's painfully clear when I use it that the responsiveness and focus is much much better.
I've the ubuntu package here - standard ubuntu compile flags, ck patchset, ready for install. If you want it and you don't want to build it yourself - I'm guessing that it should work just fine. It's worth it.
