> -----Original Message----- > From: Qingqing Zhou [mailto:[EMAIL PROTECTED] > Sent: Friday, December 16, 2005 10:13 PM > To: Dann Corbit > Cc: Tom Lane; Bruce Momjian; Luke Lonergan; Neil Conway; pgsql- > [EMAIL PROTECTED] > Subject: RE: [HACKERS] Re: Which qsort is used > > > > On Sat, 17 Dec 2005, Dann Corbit wrote: > > > > > The benchmarks say that they (order checks) are a good idea on average > > for ordered data, random data, and partly ordered data. > > > > I interpret that in linux, 5000000 seems a divide for qsortpdq. Before > that number, it wins, after that, bsd wins more. On SunOS, qsortpdq takes > the lead till the last second -- I suspect this is due to the rand() > function: > > Linux - #define RAND_MAX 2147483647 > SunOS - #define RAND_MAX 32767 > > So in SunOS, the data actually not that scattered - so more favourate for > sorted() or reversed() check?
There is a lot of variability from system to system even for the same tests. I see different results depending on whether I use GCC or Intel or MS compilers. ---------------------------(end of broadcast)--------------------------- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly