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

Home -> Community -> Usenet -> c.d.o.server -> Re: Dual processor Solaris slower than notebook Win2k

Re: Dual processor Solaris slower than notebook Win2k

From: «ó¤S°¾¤SÄF <mcc95_at_hotmail.com>
Date: 21 Mar 2001 18:03:54 -0800
Message-ID: <99bmia0u2n@drn.newsguy.com>

Actually, couple months ago I have similiar test like you and the report are pretty similiar like you have. The RAM size does NOT mean anything , because when you doing test your machine pretty much only one user in there. I found the CPU speed is most important when you doing SQl statement or exp/imp. The HD may NOT that important. Compare witH Solaris and win2000, the win2000 I/O is faster than Solaris. I think right now SUN Solaris only on stability and scalibility better than MS Win2000.

In article <3D9u6.17781$PR.148354_at_news1.wwck1.ri.home.com>, "Brian says...
>
>I have two machines running a Oracle 8.1.6 database that was created from
>the same DDL. The only thing I changed was the file naming conventions.
>Otherwise, the physical layout of the database on both machines is the same.
>
>Solaris 8
>2 processors
>1GB RAM
>Fast SCSI drive
>
>Windows 2000 Professional
>1 processor
>128MB RAM
>Slow notebook IDE drive
>
>Now, I run the import utility on both of these machines to pump 1.2 million
>rows into a table. There are no indexes or constraints on the table. The
>Solaris box takes 11 minutes to run the import and the Win2k notebook takes
>2 minutes to run the import.
>
>Here is the sql_trace of the import's INSERT statement for both machines.
>The counts are a bit different because I didn't get the trace started at
>exactly the same time, but they are close enough to show that there is a
>problem.
>
>Solaris 8
>
>call count cpu elapsed disk query current
>rows
>------- ------ -------- ---------- ---------- ---------- ---------- ------
>----
>Parse 0 0.00 0.00 0 0 0
>0
>Execute 454 28.77 434.96 0 22128 117475
>1209315
>Fetch 0 0.00 0.00 0 0 0
>0
>------- ------ -------- ---------- ---------- ---------- ---------- ------
>----
>total 454 28.77 434.96 0 22128 117475
>1209315
>
>Win2k
>
>call count cpu elapsed disk query current
>rows
>------- ------ -------- ---------- ---------- ---------- ---------- ------
>----
>Parse 0 0.00 0.00 0 0 0
>0
>Execute 441 41.56 129.25 0 8430 35083
>1203446
>Fetch 0 0.00 0.00 0 0 0
>0
>------- ------ -------- ---------- ---------- ---------- ---------- ------
>----
>total 441 41.56 129.25 0 8430 35083
>1203446
>
>I assume that the problem is that the "get buffer" counts for the Solaris
>box are so much higher than the Win2k box. What could cause this?
>
>Later,
>BEDick
>
>
>
Received on Wed Mar 21 2001 - 20:03:54 CST

Original text of this message

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