Re: why does 11g plsql_code_type default to interpreted?

From: Mark D Powell <Mark.Powell_at_eds.com>
Date: Mon, 3 Mar 2008 07:41:21 -0800 (PST)
Message-ID: <96bdf326-db77-4bc9-8115-ed995e8595d6@34g2000hsz.googlegroups.com>


On Mar 3, 9:46 am, hpuxrac <johnbhur..._at_sbcglobal.net> wrote:
> I have a test system running 11g 64 bit on OEL 4.  Created a new
> database ( custom ... not from template with pre-suppplied
> datafiles ).
>
> It seems kind of curious that oracle is defaulting plsql_code_type to
> interpreted.
>
> If the native is so much faster ... and you don't need a compiler/etc
> anymore ... why is this happening for a new database?
>
> Are there a bunch of bugs and caveats that are still being experienced
> in this area?  Maybe this default will eventually change by 11.2?

Think back to the java verse PL/SQL discussions. I think the issues are the same.

My guess as to why PL/SQL defaults to interpreted execution is becuase first this is how the code ran in the past and second in most cases the interpreted code is just as fast as compiled code. Compiled code really only has an advantage when most of the time is not spent in SQL calls, which are passed to the SQL engine. Most existing PL/SQL is not calculation intensive and spends most of its time performing SQL calls.

IMHO -- Mark D Powell -- Received on Mon Mar 03 2008 - 09:41:21 CST

Original text of this message