On Wed, 2005-02-16 at 10:04, Hagai Yaffe wrote:
> I have got an application that is currently using an in house grwon database, i am
> looking at replacing this db with MySql. I am not sure if it would be better for me to
> work with the standart MySql db or to use an embedded MySql that will be compiled with my
> application and run with it.
> Have any perfromance tests were done to copamre the stdandart MySql with an embedded
> one, I would have guss that using embedded MySql would give better perfromance, am i right
> I have used the C MySql native api to try and test the normal version and the
> embedded one, I have come acros some results that i have problem explaining. i am
> perfroming random queris on an indexed table that contains 10000 records (the table is
> indexed by it's primary key and i am querying by the indexed column). perfroming about
> 10000 repeated random queris is about 30 percent faster when using embedded MySql then
> when using normal one.
> This seems reasonable to me, but when i try to perfrom the same test on an un indexed
> table then working with the normal version is about two times faster then the embedded one
> If any one could give me any info reagrding my questions i would be greatfull.
This does not look like normal behavior.
If you could please pass us test case we will analyze the problem.
It would be best to create bug at http://bugs.mysql.com
Before doing so I however would check you're actually comparing apples
to apples - run "SHOW VARIABLES" in both versions and check if they
match each other. It could be for example in my.cnf options are only
set for standalone MySQL server so it performs faster.
Peter Zaitsev, Senior Performance Engineer
Come to hear my talk at MySQL UC 2005 http://www.mysqluc.com/
MySQL AB, www.mysql.com