Re: PostgreSQL pre-fork speedup

>>>>> "sdv" == sdv mailer <sdvmailer(at)yahoo(dot)com> writes:

sdv> Yes, I realize it's a bit old but I just wanted to make a
sdv> small point that forking is slower. It's funny you should ask
sdv> because thread creation on Linux has in fact improved over
sdv> process creation much more in 2.4 kernel.

sdv> Benchmark at IBM shows Linux 2.4 thread creation is 30x
sdv> faster than process creation. Process creation on Windows
sdv> 2000 is about twice longer than process creation on
sdv> Linux. This means forking on Win32 will be 2x slower! See
sdv> 2002 benchmark below:

sdv> http://www-106.ibm.com/developerworks/linux/library/l-rt7/?Open&t=grl,l=252,p=mgth

Excuse me for butting in here but this shows that fork AND exec is
slower than thread creation. I was under the impression that (for 2.2
or 2.4 at least) both fork and thread creation used clone (kernel not
libc). Only when a process does an exec does the diiference show
(well, actually it seems when either process modifies its stack).

Now, saying that, I have no idea how postgresql works so will shut up.

Sincerely,

Adrian Phillips

--
Who really wrote the works of William Shakespeare ?
http://www.pbs.org/wgbh/pages/frontline/shakespeare/