RE: db_file_multiblock_read_count and performance
- From: "Christian Antognini" <Christian.Antognini@xxxxxxxxxxxx>
- To: <ryan_gaffuri@xxxxxxxxxxx>
- Date: Tue, 7 Dec 2004 09:17:49 +0100
Hi Ryan
>I have been testing this extensively over the last few months. I do a full
>table scan with a
>db_file_multiblock_read_count = 1 and then one = 128( i check the 10046 trace
>to verify i am
>getting this much) and I see absolutely no difference whatsoever in response
>time.
Attached you find some results that I get during some tests that I performed on
different servers by different customers (notice that I had no influence on the
setup, I just run a test script...).
As you can see many different behaviors are to be expected.
System 1: higher values are better, of course they are "technical" limits...
(notice that 55MB/s is the maximum throughput measured on this system, i.e.
with DFMRC=32).
System 2: values higher than 16 give bad performance, i.e. the optimal value is
16.
System 3: values less than 17 are useless, i.e. at least 17 should be used to
have "correct" performance.
System 4: no performance difference was measured. Notice that this system, with
230MB/s, is also the faster I tested...
HTH
Chris
--
//www.freelists.org/webpage/oracle-l
Other related posts:
