Re: At Performance Limit for 2 GB RAM?
KimSeong-
Thanks for the response. Sorry I am a little slow in getting back to you, but rebuilding the indexes takes a while. I had already tested the index both ways you suggested originally but just to make sure I changed it back again, and still no luck in fact it seems to be about 25-50% slower than the other index.
I suspected the (stn_id, metadata_id) would perform better as well because of the lower cardinality of stn_id (4000 items) to that of metadata_id (50,000 items). But strangely the results didn't confirm this.
The y_metadatatemp table is a temporary table more for troubleshooting a more complex query so I didn't think to index it, but indexing it doesn't show any significant performance increase.
If you have any other ideas I would be interested in hearing them. I have a white paper on a Dell/MySQL benchmark that shows a database with sizes in the same ball park I have here sustaining 1900 transactions per minute with an average response time of 1.217 (s) per query. (The full details of the test are a little vague, but they lead me to believe that the times I am seeing for comparable row quantities and hardware are a bit low.)
Thanks again,
Nick
p.s. Anyone else have any input?
Subject
Views
Written By
Posted
7562
July 20, 2005 09:42AM
3327
July 20, 2005 05:45PM
Re: At Performance Limit for 2 GB RAM?
3195
July 22, 2005 01:33PM
3114
July 22, 2005 02:06PM
3501
July 22, 2005 02:13PM
2740
July 25, 2005 01:40PM
2923
July 26, 2005 03:47PM
2775
July 26, 2005 06:55PM
2879
July 26, 2005 07:06PM
2902
July 27, 2005 09:46AM
2772
July 30, 2005 11:44AM
3361
August 01, 2005 11:18AM
3326
August 04, 2005 07:17PM
3820
August 08, 2005 09:08AM
2845
August 02, 2005 12:59AM
3121
August 08, 2005 09:26AM
2732
August 09, 2005 07:57AM
Sorry, you can't reply to this topic. It has been closed.
Content reproduced on this site is the property of the respective copyright holders.
It is not reviewed in advance by Oracle and does not necessarily represent the opinion
of Oracle or any other party.