Re: Performance question.

On Wed, Mar 27, 2002 at 11:40:01AM -0800, pgsql-gen Newsgroup wrote:
> Subject: Performance question.
> From: "Wei Gao" <weigao@hotmail.com>
>  ===
> Hi, List
>
> I have a question about PostgreSQL performance.
>
> I did a test on both postgresql 7.2 and MS SQL Server 2000 by using JAVA,
> both on Windows 2000 platform. JDBC driver for postgresql come with
> postgresql, Opta2000 as MS SQL Server 2000 driver.
>
> Exact same insert operation (prepared statment), MS SQL Server with Opta2000
> use about 381 ms, and PostgreSQL use 581ms.
>
> That's about 50% performance down from MS SQL Server to PostgreSQL.

Interesting figures but any measurements under 1 second are more than likely
swamped by incidental disk activity. If you do 1000 inserts, is the result
still the same?

Also it makes a really big difference whether they are all in the same
transaction or in individual transactions.

Other than that, you have given absolutly no information about your "test".

> So, my question is:
> 1. Is this a knowing issue that PostgreSQL has a poor performance?

Your "test" above indicates no such thing. We use it exclusivly here and
it's easily fast enough.

> 2. If same hardward platform, does PostgreSQL running on Linux has a better
> performance than Windows? And how about PostgreSQL running on Linux compare
> to MS SQL Server running on Windows?

No idea.

> 3. Does JDBC Driver effect database perforce? I tried to use Microsoft JDBC
> Driver for MS SQL Server 2000, of cause, beta version. But unfortunately, I
> even can not get my program running on it.

Probably. But they're hard to get around or even measure. Simply include the
time there as part of your testing.

--
Martijn van Oosterhout <kleptog@svana.org>   http://svana.org/kleptog/
> Ignorance continues to thrive when intelligent people choose to do
> nothing.  Speaking out against censorship and ignorance is the imperative
> of all intelligent people.