MySQL Forums
Forum List  »  Performance

Re: Can we upgrade TPC-C performance test result for mysql VS Sqlserver2000
Posted by: Peter Zaitsev
Date: October 24, 2005 05:21PM

Hi george,

Your results are very interesting. There are number of things which may be optimized.
It would be best if you'd drop me an email to peter@mysql.com so we can discuss it more convenient way - I'm not that frequently checking forum to spot your reply.

In general I would first wonder why do you run the TPC-C benchmark ? Do you do it for intenral evaluation purposes or do you plan to publish it ? As I see you use innodb_flush_logs_at_trx_commit=0 which is of course specs voilation, you also use plain queries instead of stored procedures which is of course inefficient.

I would also be very interested to know which load are you getting on your server - I guess the load is disk bound ?

One of the things which might be the problem is isolation mode - you're using very strict SERIALIZABLE isolation mode. Instead you could use "SELECT FOR UPDATE" for queries which require serialization and leave default isolation mode instead.

Options: ReplyQuote


Subject
Views
Written By
Posted
Re: Can we upgrade TPC-C performance test result for mysql VS Sqlserver2000
1277
October 24, 2005 05:21PM


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.