Wade Chandler wrote:

Just wondering.  Did you know that every Postgres connection creates a
system process, and not a thread?  PostgreSQL uses Fork.  I don't use it
because of this.  DB Connection = Separate Process.  I use Firebird.
You can use Firebird 1.5 even.  That's just a consideration since you
are posting about performance.

AFAIK, a process and a thread look similar to the CPU. They are both tasks, it is just that threads share the same memory table descriptor and some other things. Sure, creating a new process implies separate copies of stack and data segments, but with CoW (Copy-on-Write) mechanism, it can save a lot of copying.


I wouldn't bet on a very high performance gain, unless the problem is more suited for the multi-threaded model (concurent tasks need a tight coupling). Is DB backend this kind of problem? It could be...

Nix.


--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to