There was an interesting presentation at PG Con from a guy at Sun who did a series of load tests on 8.3 vs 8.4
http://www.pgcon.org/2009/schedule/events/124.en.html There is a link to the video from that page so you can watch it. But he found a strange "corner case" where 8.4 performed way worse. After he did a bit of digging he found a couple of default settings that had changed in 8.4, and when he set them back to their old 8.3 values and re-ran the tests, there was a huge difference in outcome. On Wed, Jun 17, 2009 at 10:50 AM, Todd A. Cook<tc...@blackducksoftware.com> wrote: > Hi, > > First, the numbers: > > PG Version Load time pg_database_size autovac > ---------------------------------------------------------- > 8.2.13 179 min 92,807,992,820 on > 8.3.7 180 min 84,048,744,044 on (defaults) > 8.4b2 206 min 84,028,995,344 on (defaults) > 8.4b2 183 min 84,028,839,696 off > > The bulk of the data is in 16 tables, each having about 55 million rows of > the form (int, int, smallint, smallint, int, int, int). Each table has a > single partial index on one of the integer columns. The dump file was 14GB > compressed. > > The loads were all done on the same machine, with the DB going on a pair > of SATA drives in a RAID-0 stripe. The machine has 2 non-HT Xeons and > 8GB RAM. maintenance_work_mem was set to 512MB in all three cases. > > -- todd > > -- > Sent via pgsql-general mailing list (pgsql-general@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-general > -- “Don't eat anything you've ever seen advertised on TV” - Michael Pollan, author of "In Defense of Food" -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
