Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle/VMS performance problems with Multithreaded Server

Re: Oracle/VMS performance problems with Multithreaded Server

From: I.A. Saez <i.a.saez.scheihing_at_tue.nl>
Date: Wed, 08 Nov 2000 07:35:33 +0100
Message-ID: <3A08F435.5B417C7F@tue.nl>

Jan Vorbrueggen wrote:
>
> "I.A. Saez" <"i.a.saez.<secret>"@tue.nl> writes:
>
> > That MTS does not work under Unix is not such a problem. Proces creation
> > (for dedicated processes) is cheap under Unix. Process creation under
> > VMS costs a lot more (seconds!)
>
> This is mostly folklore. Actual process creation is remarkbly similar in Unix
> and VMS. What differs is that the created process by default receives a copy
> of the original process's address space in Unix; however, this isn't used that
> often (most fork() calls are followed by an exec() call). What tends to make
> the perception of process creation as slow is the additional overhead of
> setting up the environment of DCL (symbols, logical names) when SPAWNing (when

Thus process creation on VMS is slower:-)

> will that improvement to the DCL SPAWN code finally be released, I wonder!?);
> however, all that can be turned off. And on a current Alpha, the situation is
> quite different fron a VAX-11/725 with regard to the resources consumed by
> a $CREPRC call.
>
> Jan
Received on Wed Nov 08 2000 - 00:35:33 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US