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

Home -> Community -> Mailing Lists -> Oracle-L -> RE: db_file_multiblock_read_count and performance

RE: db_file_multiblock_read_count and performance

From: <ryan_gaffuri_at_comcast.net>
Date: Mon, 06 Dec 2004 19:50:36 +0000
Message-Id: <120620041950.14292.41B4B80C0005CCAB000037D42207002953079D9A00000E09A1020E979D@comcast.net>


i know how the CBO works. I've read the article. Per Oracle docs, they lead you to believe, that oracle will do Physicall IO more efficiently if you set that parameter higher and your system can use it. This does not appear to be true. Here is how the test cases work.

1. determine max db_file_multiblock_read_count
2. bounce the instance to flush all the cache.
3.query with a full table scan with the value set = 1
4. bounce the instance to flush all the buffers
5. query with a full table scan with the value set to the max for my system. 

No difference. I know query plans change with this setting. However, the documentation appears to be a bit misleading lead me to believe this alone would improve response time other than just tinker with the the CBO estimates.

> Do you get a different plan when you change the parameter?
>
> Are you doing any physical IO with your test query?
>
> I thought your OS will always do whatever it is capable of, so changing
> this parameter will only effect the cost of the different plans Oracle
> comes up with.

--
http://www.freelists.org/webpage/oracle-l
Received on Mon Dec 06 2004 - 13:55:42 CST

Original text of this message

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