HPUX vs Windoze

[jluckhoff]

Hi

Ran the following query locally on HPUX 8.1.6 machine:

*****************************************************************

SQL> alter system flush shared_pool;

System altered.

SQL> set timing on
SQL> alter session set timed_statistics=true;

Session altered.

Elapsed: 00:00:00.01
SQL> set autotrace traceonly
SP2-0613: Unable to verify PLAN_TABLE format or existence
SP2-0611: Error enabling EXPLAIN report
select * from big_table;
set autotrace offSQL>

685967 rows selected.

Elapsed: 00:07:37.73

Statistics
----------------------------------------------------------
276 recursive calls
6 db block gets
68392 consistent gets
24225 physical reads
0 redo size
163527679 bytes sent via SQL*Net to client
5076570 bytes received via SQL*Net from client
45733 SQL*Net roundtrips to/from client
4 sorts (memory)
0 sorts (disk)
685967 rows processed
SQL> alter system flush shared_pool;

*********************************************************************

As you can see the query took 7:37 minutes.

I then ran the same query on a windoze client connecting to this same HPUX install

Here is the result:

******************************************************************

SQL> alter system flush shared_pool;

System altered.

SQL> set timing on
SQL> alter session set timed_statistics=true;

Session altered.

Elapsed: 00:00:00.00
SQL> set autotrace traceonly
SP2-0613: Unable to verify PLAN_TABLE format or existence
SP2-0611: Error enabling EXPLAIN report
SQL> select * from big_table;

685967 rows selected.

Elapsed: 00:04:272.22

Statistics
----------------------------------------------------------
276 recursive calls
6 db block gets
68392 consistent gets
24225 physical reads
0 redo size
125403279 bytes sent via SQL*Net to client
3064287 bytes received via SQL*Net from client
45733 SQL*Net roundtrips to/from client
4 sorts (memory)
0 sorts (disk)
685967 rows processed

SQL> set autotrace off

******************************************************************

As you can see this test ran almost 3 minutes faster:

Why? - Surely if running the query locally on the HPUX machine it should run as fast or faster than running it from a windows client machine.

thanks

j

[TomazZ]

If I understand correctly both time is the same database instance, you changed only client.
Both time should run same provided you have same circumstances.
Since alter system flush shared_pool does not clear the buffer cache, you should repeat the test with restarting the database instead.

[abhaysk]

Also Note

HP UX :

163527679 bytes sent via SQL*Net to client
5076570 bytes received via SQL*Net from client


WIN :

125403279 bytes sent via SQL*Net to client
3064287 bytes received via SQL*Net from client


The data sent/received on n/w path was 1.7 times more in ur HPUX clinet machine than WIN client.