Re: RBTree implementation and thread structure
"Salvatore Benedetto" <[email protected]> Sun, 13 Apr 2008 08:50:50 +0000
| Newsgroups | gmane.os.openbeos.kernel.devel |
|---|---|
| Message-ID | <[email protected]> |
Hi Andrè, nice to hear you are still around! On 13/04/2008, André Braga <[email protected]> wrote: > On Sat, Apr 12, 2008 at 6:43 PM, Salvatore Benedetto <[email protected]> wrote: > > Hi there, > > I have two quick questions. > > First: do we have a red and black tree class implementation that I > > could use in the scheduler? Just wondering, not actually doing any > > coding yet. > > I see we have a AVLTreeMap class implemented by Ingo, but looks like > > it's not used anywhere. > > > Uhm, I guess I'm supposed to give a sign of life so that we don't step > on each other's toes... :) > > FWIW, I have *not* abandoned the scheduler I wrote for the '07 GSoC; > on the contrary, I'm working on it daily as part of my final > assignment as a college student. Since things are pretty much in flux > now that I'm being oriented by the professor that teaches the > Operating Systems class (i.e., lots of things have changed because of > his feedback), the code isn't really in a shape that I can integrate > into the Haiku tree. But I'm evidently getting there :) Glad to hear it. > > BTW, this is just a "ping" message. Salvatore, if you really want to > work on the scheduler for your own enjoyment, be my guest :) But only > for that reason. My "baby" is taking shape and I'm going to feel 10 > times more stupid than I currently am if you step up and put your own > effort into it, given that I was actually PAID to do this. > > I indeed had that period of excess of real life issues, but I'm on > much more solid ground now, having finally regained faith in myself. > > Anyway, I'm probably going to post a blog entry discussing the new > scheduler model anytime soon, since the one I was using fails in > terrible ways on real life situations. And I'll be the first to admit, > even though I took a completely different path to reach those results, > the current model resembles the infamous CFS a lot. > > Stay tuned :) First of all, I'm very sorry for your real life issues. Hope everything is over now. As far as I'm concerned though, I am too taking a class about Operating Systems and parallel programming, and that's why I'm trying to work on it. I didn't know you were still around working on it. Anyway, I think it is not that bad if we work on two different schedulers for the same OS, as we could benchmark both and see the one that fits better, or better yet, we could "steal" each other ideas to make the best out of it. Competition always brings good results. Anyway though, scheduler it is not the first on my priority list, OHCI is. I've read all your blog posts of course, they have been very helpful. I'm looking forward to read your next status update on this project. Do me a favor though, this time keep it short and cut out the unnecessary stuff! ;-) > > > Cheers, > > A. > > Thanks, -- Salvatore Benedetto (a.k.a. emitrax) Student of Computer Engineer University of Pisa www.haiku-os.it ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone