2005 any faster than 2000?

[bytehd]

Anyone do some benching?

George Walkey
Senior Geek in charge

http://www.insyncva.com

 

[mrdenny]

Not yet, but shortly. I'll post when I've got some info. I've got a 2005 cluster that we are installing tomorrow and a 2000 cluster that are both on identical hardware so I should be able to get some decent info.

Denny
MCSA (2003) / MCDBA (SQL 2000)

--Anything is possible. All it takes is a little research. (Me)

http://www.mrdenny.com

 

[chiph]

bytehd -
What criteria is important to you? Inserts/updates vs. selects vs. analysis cubes?

Chip H.


____________________________________________________________________
Donate to Katrina relief:

http://s1.amazon.com/paypage/PELYGQVJ8Q7IB/103-6821258-5919825

If you want to get the best response to a question, please read FAQ222-2244 first

 

[SQLDenis]

Here is a link for you (

http://www.sqljunkies.com/Article/077C7BEB-EB31-4A07-923D-BE309F59D0F8.scuk)

Denis The SQL Menace
SQL blog:

http://sqlservercode.blogspot.com/

Personal Blog:

http://otherthingsnow.blogspot.com/

 

[bytehd]

Chiph:
for me querying speed
but inserts are nice too
i KNOW SSIS is faster than DTS



George Walkey
Senior Geek in charge

http://www.insyncva.com

 

[mrdenny]

Here's my config. I've got two, two node clusters setup.

Each are 4 proc servers running Windows 2003 AS.
Cluster1 is running SQL 2005
Cluster2 is running SQL 2000
All 4 nodes have 4 GB of RAM
All 4 nodes have 3.33 Ghz Intel Processors with HT enabled
AWE settings are the same on both servers, and both servers are allowed to use up to 3515 MB of RAM.
SQL Storage is on a SAN.
All 4 nodes are identical servers and were purchased within 1 week of each other.

Here is what I saw:

First I was inserting 1000 records one at a time, then deleting 1000 records one at a time. Each server was allowed to run for 31.5 minutes.
SQL 2000 - 454,018 Records in 31.5 minutes
SQL 2005 - 586,496 Records in 31.5 minutes

SQL 2005 was able to insert and delete an additional 132,478 records.

I then ran a basic select statement to see how the execution plans on each server would function. The query was:
select count(*), min(d), max(d), avg(d)
from Table1

SQL 2000
Number of Selects: 3
Rows returned by Selects: 5
Total Execution Time: 31 ms

SQL 2005
Number of Selects: 2
Rows returned by selects: 2
Total Execution Time: 15 ms

CPU Speed:
SQL 2000 - 10-11%
SQL 2005 - 10-14%

While these are some basic results they are interesting.

Denny
MCSA (2003) / MCDBA (SQL 2000)

--Anything is possible. All it takes is a little research. (Me)

http://www.mrdenny.com