Hi everybody, First of all, before anybody attacks me, I do not intend to start a flamewar. I have simply been testing out SQLObject and SQLAlchemy/Elixir for a program I'm writing (I previously posted on some issues I had). If I am to be honest, I like SQLObject a lot more, because its syntax and way of doing things is, IMHO, simpler and more Python-like. And I wouldn't have tried out Alchemy in the first place if it wasn't for Elixir.
After battling between tutorials and documentation, I finally managed to create another development branch, based on my SQLObject one, fully working with Alchemy/Elixir. I did this (keeping two copies of the program in the same repo, with different backends) so as to do a speed comparison between the two. Again, not that I don't like SQLO, but I just wanted to see which one was faster, since it benefits my application, and WHY it was faster. The app I am talking about is a filesystem indexer, like Gnomecatalog or Basenji. Basically it scans the filesystem recursively and for every entry (file or directory), it extracts some metadata about it and stores it in a DB, so it can be accessed later (perfect for those storage HDs you have full of stuff you don't want to index by hand :)). Originally I just kept the whole scan content in memory and pickled it when asked, but with huge directories it turned into a problem, so I started looking for an alternative in the DB field. I haven't thoroughly scanned it, neither is my scanning method the most scientific one, but the results I got so far seem to point out that Alchemy/Elixir is faster than SQLO. I did two scans, one with some small text files in just one directory, and another one with plenty (1GB+) of wallpapers split into two directories. Here are the results: SQLObject: Text files: real 0m25.704s user 0m3.240s sys 0m1.424s Wallpapers: real 5m25.644s user 3m6.760s sys 0m16.881s SQLAlchemy/Elixir: Text files: real 0m13.754s user 0m6.696s sys 0m0.504s Wallpapers: real 3m58.949s user 3m4.392s sys 0m7.244s Of course, numbers may not be that big, but there is a difference which I figure would increase on bigger scans. Are there any ideas on why would SQLO be a bit slower than Alchemy? Is there something that can be done to amend this? If anybody wants to check out the code, its at git://vicarious.com.ar/indexor.git, or at github on http://github.com/godlike64/indexor. The branches are dev/godlike/elixir for the Elixir one, and dev/godlike/bleeding for the SQLObject one. Any questions feel free to ask me :) Cheers Juan Manuel Santos PS: the program is of course in development, so some things may not quite work. Actually, I timed it by opening, ordering it to scan a directory, and closing the window. After the scan finishes (in complete background), the program exits and you get the time prompt). ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ sqlobject-discuss mailing list sqlobject-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sqlobject-discuss
