From: Carsten H. (T. R. <ra...@ra...> - 2005-09-25 02:34:23
|
On Sat, 24 Sep 2005 03:10:12 +1000 David Seikel <won...@ya...> babbled: > On Sat, 24 Sep 2005 00:33:07 +0900 Carsten Haitzler (The Rasterman) > <ra...@ra...> wrote: > > > On Fri, 23 Sep 2005 22:26:44 +1000 David Seikel > > <won...@ya...> babbled: > > > > > takes seventeen seconds to boot on a test box that is one tenth the > > > power of my Athlon. > > > > 17 seconds! you can do better than that! i challeng you to get it to > > below 10! i am actually quite sure it can be done. i've had a machine > > boot lilo ->X running glxgears in 1.6 seconds. (yes linux) there are > > well - come hacks invovled, but you could manage 10 to a graphical > > login i would say. :) > > The 17 seconds is for an average Pentium 300, I'm sure my Athlon 3000+ > can do a lot better. I'm waiting for Xorg to release their modular X, > which should happen next month according to their web site. Then I can > apply my 'start the bare minimum to support the user logging in and > doing something useful, then start the rest in the background' tricks to > X. At the moment, the bulk of getting logged in graphically is spent in > X before it starts the display manager. actually - here's a hint. x actually loads EVERY module when it starts to find out whats there - that means every driver, every extension file etc. it may only end up using a few but it literally does a heap of disk work on them anyway (for interest - start x and put it under strace - log it and check the output and see all the disk io it does for modules etc.). one great way to get x to start fast - remove anything you dont use or need :) (or fix x to not go read everything!) :) another is to hack the driver to to avoid useless re-setting of chips (eg in laptops often you can get another 0.5 seconds by not resetting the gfx controller - the backlight stays on and you get an almost seamless transition to gfx mode. this of course could have other implications... like on buggy chipsets or ones with unknown states it may not work properly) > > but but! congrats on CARING and WORKIng at it. it seems this is not > > somehting major distros care much about and i agree is a pain. slow > > boots. bloat. everything. agreed 100%. > > I thought you would appreciate it. I'll keep you informed. > > > i am so glad you brought this up. this now enters one of my favorite > > domains: OPTIMISING! i love optimising things :) and well - i didnt > > realise a large collection of eaps would kill you. for info - how many > > do you have? just give me a quick count of /bin/ls > > ~/.e/e/applications/*.eap | wc -l > > I also love optimising things, I have done a lot of work with > micro controller powered appliances. > > 1641 eap files. To put that into perspective, SuSE 9.3 Professional > ships with 4061 packages for x86, although some of those are libraries, > fonts, and other things that don't end up in the menu. hmm - i still think you have some gross disk IO issues. i may have only a fraction - but e starts in like < 1 second for me (200 not 1600, but i dont get like a 20 / 8 (2.something) second extra lag. what fs are u using? ext2? ext3? xfs? reiser? what disk? whats its IO capacity? (hpardm -t /dev/device) ? > > of the eap colelciton a nightmare. the solution i'm thinking now is a > > cache - a cache file listing all metadata for all .eap's in a dir that > > can be updated by E itself or an external cmd-line tool. e will load > > the cache first (ine one big seamless load from start to end) and then > > in slow-burn timers scan the cache to match up real files. if real > > fiels have changed since the cache was made, or if files have vanished > > they will be remvoed from the cache and the cache re-written. files > > not in the cache will be added. this should improve things by light > > years. :) > > Sounds good to me, and I'll be glad to be the guinea pig for that, or at > least send you my eap collection. Just be careful to get the cache if i need to i can fake up a massive collection :) but i'd prefer a small one for working and ebugging/dev purposes for now :) btw - partt way through. my 200 eap's use about 6mb of disk. the cache file for that dir is a touch under 7kb. thats all e has to load on startup - so it should be massively faster for you in the future. i'm still workign on it and its not enabled in production code yet. :) > checking working well and robust, I have seen some shocking > implementations of similar ideas. I spend a lot of my time these days trust me. if there;'s something i've done a LOT it's caches! :) one reason e is fast IS its caching. evas caches, edje caches... they have in memory caches to avoid re-opening files and decoding them from disk. this avoids an io pass and a decode - but even so theres still a kernel disk cache behind that too :) caches are good. caches RULE! :) > coding in Java, and it's a pet peeve of mine that I have yet to see a > properly working Java cache implementation that doesn't require me to > just delete it on a regular basis. By regular, I mean every time I > recompile my Java app, which can happen once a minute during heavy > debugging sessions. I have a Java cache delete built into my build tool > now. trust me. so far e is building and maintaing its own cache. well it builds - it just wont modify it once built atm - i need to fix all this. never fear. working on it. > While on the subject of excessive eap collections, managing 1641 eaps > with entangle using a thin, scrollable strip of them is a real pain. Is > it possible for a theme to make that a big, wide area for the eaps? > Smaller icons? Organise them in the same way that e17genmenu organises > them (by subject)? Filter them? All of the above? > > It would be a logical step, and probably quite easy to implement, to be > able to edit an eap from within entangle by just clicking on the icon. > > Oh well, back to struggling to get a cvs update of e17 out of > sourceforge. Man that can be painful some times. Yesterday I did a cvs > update out of sourceforge for ONE OF MY OWN PROJECTS, using my account > and everything. Took five hours for something that should have taken > minutes. Hmm, got e17, but can't get the misc module yet. i havent seen anything that bad yet. are you pn dialup? :) > -- > Stuff I have no control over could be added after this line. > > Send instant messages to your online friends http://au.messenger.yahoo.com > > > > ------------------------------------------------------- > SF.Net email is sponsored by: > Tame your development challenges with Apache's Geronimo App Server. Download > it for free - -and be entered to win a 42" plasma tv or your very own > Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php > _______________________________________________ > enlightenment-devel mailing list > enl...@li... > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ra...@ra... 裸好多 ra...@de... Tokyo, Japan (東京 日本) |