Constructing a Database Benchmark (average load)

Our first benchmark was custom-written in .NET, using ADO.NET to connect to the database. The AnandTech Forums database, which is over 14GB in size at the time of the benchmark, was used as the source database. We'll dub this benchmark tool "SQL Loader" for the purposes of discussing what it does.

SQL Loader allows us to specify the following: an XML based workload file for the test, how long the test should run, and how many threads it should use with which to load the database. The XML workload file contains queries that we want executed against the database, and some random ID generator queries that populate a memory resident array with ID's to be used in conjunction with our workload queries. The purpose of using random ID's is to keep the test as real-world as possible by selecting random data. This test should give us a lot of room for growth, as the workload can be whatever we want in future tests.

Example workload:

< workload>

< !--- A SAMPLE WORKLOAD QUERY THAT RETURNS ALL THE FIELDS FROM THE PRIVATEMESSAGES TABLE RANDOMLY --->

<query>

<code>select * from privatemessages where imessageid = @pmessageid</code>

<type>read</type>

<randkey>pmessageid</randkey>

</query>

<!--- RANDOM ID GENERATOR FOR SELECTING RANDOM PRIVATE MESSAGES --->

<randomid>

<rcode>select imessageid,newid() as pmsgid from privatemessages order by pmsgid</rcode>

<name>pmessageid</name>

</randomid>

< /workload>


A screenshot of the SQL Loader


Test Information

The workload used for the test was based on every day use of the Forums, which are running FuseTalk. We took the most popular queries and put them in the workload. Functions, such as reading threads and messages, getting user information, inserting threads and messages, and reading private messages, were in the spotlight. Each iteration of the test was run for 10 minutes, with the first being from a cold boot. SQL was restarted in between each test that was run consecutively.

The importance of this test is that it is as real world as you can get; for us, the performance in this test directly influences what upgrade decisions we make for our own IT infrastructure.

Hyper Threading and The Tests AnandTech Forums Database Test Results
Comments Locked

44 Comments

View All Comments

  • RyanVM - Monday, September 13, 2004 - link

    #9, why would they be in any rush whatsoever to do that when DDR2 is slower and more expensive clock for clock? Besides, with their on-die memory controller, memory performance is far from a limiting factor for K8 systems (look how little difference there is in performance between single channel and dual channel memory for K8 systems).
  • karlreading - Monday, September 13, 2004 - link

    Hmm, seems that the nocona has certainly helped xeons save face against K8. with iAMD64 and 800Mhz FSB they do look much healthier compared to prestona Xeon, raising the perfromance bar and giving the intel fans ( and Dell!! LOL ) the 64 bit option that they always deserved, rather than the expensive, unpopular niche itanic . Considering the implication of having the two CPU's fighting each other for memory bandwidth, something on netburst parts cant get enougth of, the xeons keep pace superbly and it makes one wonder what they would be capable of if each nocona had its own dedicated memory banks and controller!!! Heres one AMD fan hoping opterons @ 2.6Ghz ( opteron 260??) come out quick to keep the chipzilla sweating!!!
  • Viditor - Monday, September 13, 2004 - link

    Excellent first (32bit) look, Jason and Ross! Well written, Many thanks!
    I look forward to the next edition as well.
  • Jason Clark - Monday, September 13, 2004 - link

    Mino, actually AMD sent the tyan and 250's together :) So, they must have some faith in the board since they knew exactly the purpose of it. Aside from that the 250 managed to scale very very well, so I doubt that the board is any issue what so ever... The 64 bit article will be based around 64 bit windows.. not linux. Linux stuff is handled by kris in the linux section.

    Cheers
  • mino - Monday, September 13, 2004 - link

    by "BEST compiler around" I off course meant by performance, read ability to use the hardware given most efficiently (in case of x86-64 it doesn'nt have to be stable, stability will come by time, these chips will be there for long time).

    Just a little expl. ;0
  • mino - Monday, September 13, 2004 - link

    hmm, Jason:

    Please could You explain why with Nocona You do consider it "obviously used the E7520 Lindenhurst server chipset" and with Opteron it is allright to use purely Workstation chipset/board ???

    I know there should just minor difference betwen K8W and K8S Pro , but this "minor" rise in K8 performance might have changet this very narrow race !!

    Other than That I'm waiting for 64-bit comparison, PLEASE do include 32bit results in there then. I would preffer 32bit Nocona optimized code vs 64bit Nocona optimized vs 32bit K7 optimized(runs better than K8 optim.) vs AMD-64 optimized(compiled with BEST compiler around-even if it will be only beta compiler used) code on Opteron. This should be done on linux with systems compiled from scratch for this test. I know This is much of work, but there are people around that can make such a system up and running in 3 to 4 hours.(one is sitting near me:). So, IF You need help, call some friends of yours. Since such an comparison would be ONE and the BEST ONE around this pitty planet.

    good luck,
    mino
  • jrphoenix - Monday, September 13, 2004 - link

    #12... Very funny.... I had to go back and read it again. I guess they did fart out the specs, LMAO!!!
  • Jason Clark - Monday, September 13, 2004 - link

    A 3.2 prestonia vs. 3.2 nocona could be interesting though, I'll see what we can do there.
  • Jason Clark - Monday, September 13, 2004 - link

    We are in planning a 64bit db test, and web test. Not to worry. These types of articles that use real world tests, take time :)
  • Cusqueno - Monday, September 13, 2004 - link

    I believe the authors meant "piqued their interest" instead of "peaked their interest". I don't think this is a reference to climbing mountains.

Log in

Don't have an account? Sign up now