On 1.1.2013 17:35, Alvaro Herrera wrote: > There was an earlier suggestion by Andres Freund to use memcmp() > instead, but I don't see that in the latest posted version of the patch; > was there a specific rationale for taking it out or it was just lost in > the shuffle?
No, I've tried that approach with a comparator like this: static int rnode_comparator(const void * p1, const void * p2) { return memcmp(p1, p2, sizeof(RelFileNode)); } but it turned out to be slower than the current comparator. I've posted some benchmark results and possible explanation on 20/12 (message 50d26fe8.1040...@fuzzy.cz). If you could verify my results, that'd be great. Tomas -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers