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

Home -> Community -> Mailing Lists -> Oracle-L -> High elapsed to CPU ratio for parses

High elapsed to CPU ratio for parses

From: Paul Baumgartel <paul.baumgartel_at_gmail.com>
Date: Mon, 22 Aug 2005 18:01:22 -0400
Message-ID: <f8c4771050822150154ecf2ad@mail.gmail.com>


I don't think I've seen this particular problem before. Oracle 10.1.0.4.0 on Windows.

While database is under heavy load as part of a performance test (using Mercury LoadRunner), a number of SQL statements (with bind variables, no literals; have been part of the application for a long time) are exhibiting multi-second elapsed time for a single parse, with CPU time for the parse relatively high (0.35 sec) but still far below the elapsed (6.60 sec for this example).

TKPROF output of 10046 trace does show some waits for library cache and shared pool, but not nearly adding up to the times shown above. I'm at a bit of a loss to know where to look for the consumer of 6 seconds for a single parse of a single SQL. Any suggestions welcomed.

-- 
Paul Baumgartel
paul.baumgartel_at_aya.yale.edu
--
http://www.freelists.org/webpage/oracle-l
Received on Mon Aug 22 2005 - 17:03:37 CDT

Original text of this message

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