Tag: performance
-
Performance vs usability
I had an idea to make Evolution display more human-readable/stuff-I-care-about-only sender information, but it was scrapped. It is sad when you hit technical roadblocks in technology (here, performance considerations) that trump usability (well, that and the fact that the devs were not too enthusiastic about the idea to begin with).
-
My browser needs a nuclear-equipped walking death-mobile database
Who thought that using a SQLite database (what I refer to as a nuclear-equipped walking death-mobile) to manage everything (including bookmarks) in Firefox (or other apps) was a clever idea? A friend of mine was asking me to investigate why Firefox alwas hung when quitting on his Dellbuntu today. After an hour, I found out…
-
Import performance improvements
Sometimes I think developers should use netbooks as their testing machines. Then they would be hit by performance bottlenecks in much more obvious ways 🙂 This morning, Ed pushed a commit that limits the thumbnail size to 96×96 (previously unlimited) in PiTiVi’s master branch. What this means is that previously, each time you imported a…
-
Improved Specto startup times
After my post on profiling Specto’s startup, Wout Clymans put on his Hero Hacker Hat and fixed the major problem that was causing Specto to do too much work for nothing. Now, Specto starts in 2 seconds instead of 10 seconds. What happened there? Well, for making sure that we are using the same “internal”…
-
Profiling Specto (and whole Python applications in general)
A few months ago (when we still thought we were about to release 0.3 “real soon now” ;), I noticed that Specto is noticeably slow to start up, even on warm starts (when it is not the first time you launch it). It always takes at least 6 seconds to paint the list of watches…
-
Rhythmbox: sois patient
Ma librairie musicale: 11200 morceaux Démarrage à chaud de Rhythmbox si on le laisse démarrer «tranquillement» (sans le forcer à jouer de la musique immédiatement): 13 secondes jusqu’au moment où il n’y a plus de grattage de disque dur et de barre de progression. Démarrage à chaud de Rhythmbox si on lui demande de jouer…
-
Recent Documents
Did you know? Not clearing your recent documents periodically in GNOME means your whole desktop responsiveness suffers. Ugliest performance bug. Ever. On my laptop, ~/.recently-used.xbel weighted 4.8 MiB, and totalled 99 928 lines! In this file, I had entries dating back 2007-04-03; that’s just about one year of useless history. I will now be able…
-
Readahead personnalisé pour réduire le temps de login
Je suis un maniaque assoiffé de performances. J’ai lu des discussions, wikis, écoutés toutes les conférences d’optimisation de gnome, et j’ai peut-être bloggé plusieurs fois à propos de mes nombreuses tentatives pour presser chaque seconde de vitesse de login possible. Mais depuis six mois, j’étais un peu déprimé de constater que peu importe ce que…
-
Régression de performances de login avec compiz fusion
Mon ordinateur portatif nécessitait 45 secondes pour le processus de «login» à partir d’un démarrage «à froid» de l’ordinateur. C’est indécent me dis-je, «c’est des performances que j’attendrais d’une vieille installation Windows XP infectée». Puis j’ai désactivé compiz (activé par défaut dans ubuntu 7.10), redémarré l’ordinateur pour refaire mon chronométrage à froid, et cette fois…