Hi Francesc, hi Ümit, Il 20/01/2012 15:16, Francesc Alted ha scritto: > 2012/1/20 Ümit Seren <uemit.se...@gmail.com> > >> So I played around a little bit futher. >> I tried to use my new solution (using table.append instead of >> row.append) with following settings in table.openFile() >> METADATA_CACHE_SIZE=2*1024*1024 >> NODE_CACHE_SLOTS=1024 >> >> I saw the same performance problems like in the previous code. So it >> slows down after a while to 1 table/sec instead of 10 tables/sec >> >> When i change it to NODE_CACHE_SLOTS=0 I don't have any performance >> problems. >> >> > Ah, that's great. Apparently 1024 nodes is a bit too much for the LRU > cache, specially when nodes are not reused a lot. Hmm, perhaps it would be > a good thing to reduce this figure to something like 256 or less (some > benchmarks needed). > > Thanks for your double check!
Thank you very much Francesc and Ümit. I have filed a new ticket on github [1] for this issue. Please Ümit, if you have updated code for benchmarking I would appreciate a lot if you could contribute it. [1] https://github.com/PyTables/PyTables/issues/121 ciao -- Antonio Valentino ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users
