Re: [HACKERS] PostgreSQL Benchmarks

>Hrm. I just saw that the PHP ADODB guy just published a bunch of database
>benchmarks. It's fairly evident to me that benchmarking PostgreSQL on
>Win32 isn't really fair:

>http://php.weblogs.com/oracle_mysql_performance

And why is the highly advocated transaction capable MySQL 4 not tested?
That's the problem, for every performance test they choose ISAM tables, and
when transactions are mentioned it's said "MySQL has transactions". But why
no benchmarks?

Regards,
Mario Weilguni

On Tue, 2003-02-11 at 08:31, Mario Weilguni wrote:
> >Hrm. I just saw that the PHP ADODB guy just published a bunch of database
> >benchmarks. It's fairly evident to me that benchmarking PostgreSQL on
> >Win32 isn't really fair:
>
> >http://php.weblogs.com/oracle_mysql_performance
>
> And why is the highly advocated transaction capable MySQL 4 not tested?
> That's the problem, for every performance test they choose ISAM tables, and
> when transactions are mentioned it's said "MySQL has transactions". But why
> no benchmarks?
>

Insert Statement

Not using bind variables (MySQL and Oracle):
$DB->BeginTrans();

Using bind variables:
$DB->BeginTrans();

PL/SQL Insert Benchmark
Appears to not initiate a transaction. I'm assuming this is because
it's implicitly within a transaction? Oddly enough, I am seeing
explicit commits here.

It appears that the benchmarks are attempting to use transactions,
however, I have no idea if MySQL's HEAP supports them. For all I know,
transactions are being silently ignored.

Regards,

--
Greg Copeland <greg(at)copelandconsulting(dot)net>
Copeland Computer Consulting

On Tuesday 11 Feb 2003 8:01 pm, Mario Weilguni wrote:
> >Hrm. I just saw that the PHP ADODB guy just published a bunch of database
> >benchmarks. It's fairly evident to me that benchmarking PostgreSQL on
> >Win32 isn't really fair:
> >
> >http://php.weblogs.com/oracle_mysql_performance
>
> And why is the highly advocated transaction capable MySQL 4 not tested?
> That's the problem, for every performance test they choose ISAM tables, and
> when transactions are mentioned it's said "MySQL has transactions". But why
> no benchmarks?

I did benchmark mysql/postgresql/oracle sometime back. Mysql with transaction
is 90% as fast as postgresql. But it dies down with increased number of users
no matter how much resources you throw at it.

Oracle is 130% of postgresql. This was postgresql 7.2.x series so things have
changed for sure, but you got the idea, right?

Shridhar

There's "The Open Source Database Benchmark",
http://osdb.sourceforge.net/.

Anyone tried to use it?

__________________________________________________
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day
http://shopping.yahoo.com