Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Postgresql mentioned on Newsforge MySQL article

0 views
Skip to first unread message

Tony Caduto

unread,
Aug 28, 2006, 4:32:51 PM8/28/06
to
http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml

Don't know the validity of this dvd order test they did, but the article
claims Postgresql only did 120 OPM.
Seems a little fishy to me.

--
Tony Caduto
AM Software Design
http://www.amsoftwaredesign.com
Home of PG Lightning Admin for Postgresql
Your best bet for Postgresql Administration


---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

Joshua D. Drake

unread,
Aug 28, 2006, 5:02:12 PM8/28/06
to
Tony Caduto wrote:
> http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
>
> Don't know the validity of this dvd order test they did, but the article
> claims Postgresql only did 120 OPM.
> Seems a little fishy to me.
>

This has got to be a complete joke.

Joshua D. Drake

--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?

http://archives.postgresql.org

Peter Eisentraut

unread,
Aug 28, 2006, 5:04:42 PM8/28/06
to
Tony Caduto wrote:
> http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
>
> Don't know the validity of this dvd order test they did, but the
> article claims Postgresql only did 120 OPM.

The contest evaluated the solutions sent in by whoever wanted to
participate. This doesn't prove anything, except that the MySQL
marketing department is exceptionally slow. These contest results were
published months ago.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

Chris Mair

unread,
Aug 28, 2006, 5:05:57 PM8/28/06
to

> http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
>
> Don't know the validity of this dvd order test they did, but the article
> claims Postgresql only did 120 OPM.
> Seems a little fishy to me.

There was just one submission for PostgreSQL made by one guy who didn't
manage to finish it (c't says connection pooling didn't work).

See: http://archives.postgresql.org/pgsql-advocacy/2006-06/msg00052.php

MySQL's winning submission was by a team from MySQL AB.

Enough said...

Bye,
Chris.

PS: this is, by the way a few months old, I'm wondering why MySQL
does the press release only now...


--

Chris Mair
http://www.1006.org

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majo...@postgresql.org so that your
message can get through to the mailing list cleanly

Bill Moran

unread,
Aug 28, 2006, 5:09:26 PM8/28/06
to
In response to Chris Mair <chris...@1006.org>:

>
> > http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
> >
> > Don't know the validity of this dvd order test they did, but the article
> > claims Postgresql only did 120 OPM.
> > Seems a little fishy to me.
>
> There was just one submission for PostgreSQL made by one guy who didn't
> manage to finish it (c't says connection pooling didn't work).
>
> See: http://archives.postgresql.org/pgsql-advocacy/2006-06/msg00052.php
>
> MySQL's winning submission was by a team from MySQL AB.

It's just kind of a shame that we didn't have a chance to put together
a serious entry ...

This is the first I've heard of the contest.

--
Bill Moran
Collaborative Fusion Inc.

---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faq

Brandon Aiken

unread,
Aug 28, 2006, 5:10:21 PM8/28/06
to
Looks like it was a design contest not a benchmark to me. Surprise,
surprise, the team that personally designs a DBMS has the best
performing DBMS. The second place winner, Alexander Burger, is the
author of the solution he used: Pico LISP. The third place team,
MonetDB, used their solution, MonetDB.

The results tell me that people who write databases can write good apps
that use those same databases. And the world collectively says "O
RLY?".

--
Brandon Aiken
CS/IT Systems Engineer

http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend

Ned Lilly

unread,
Aug 28, 2006, 5:49:02 PM8/28/06
to

Chris Mair wrote:
>
> PS: this is, by the way a few months old, I'm wondering why MySQL
> does the press release only now...

Because they don't have anything else to talk about, and are filling a vacuum?

Cheers,
NL

Scott Marlowe

unread,
Aug 28, 2006, 6:06:45 PM8/28/06
to
On Mon, 2006-08-28 at 16:02, Joshua D. Drake wrote:
> Tony Caduto wrote:
> > http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
> >
> > Don't know the validity of this dvd order test they did, but the article
> > claims Postgresql only did 120 OPM.
> > Seems a little fishy to me.
> >
>
> This has got to be a complete joke.

I'm wondering if the source code is available.

My guess is it was full of MySQLisms and the postgresql "port" was
written without indexes, no transactions, and relied on running dozens
of queries that postgresql could have more efficiently done in one query
or in one transaction at least.

I wonder if one couldn't get ahold of the code and "fix it" to see what
postgresql could actually do.

We have the same problem at work. RT is a great little program that
supports several database backends. But it's dumbed down to make MySQL
3.23 happy and fast, so it's kinda a dog on PostgreSQL, even though it's
ultra stable. Seems the newest release is finally catching up on
PostgreSQL.

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

Markus Schiltknecht

unread,
Aug 28, 2006, 6:06:54 PM8/28/06
to
Tony Caduto wrote:
> http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
>
> Don't know the validity of this dvd order test they did, but the article
> claims Postgresql only did 120 OPM.
> Seems a little fishy to me.

Now, this article really s**ks! First of all, the original contest was
specifically not only about performance. And the MySQL team did a whole
lot of dirty tricks (i.e. using memcached) to push their solution.

I am the one who has written he only PostgreSQL entry, for which I'm
still sorry and ashamed, because it performs so poorly. I just didn't
have much spare time to spend, but thought I'd send it in anyway. One of
the reasons it did not perform well was, that I simply have forgotten to
enable connection pooling.

Another entry using Perl and PostgreSQL from somebody who put a lot more
time and effort into it was disqualified because the benchmarker claimed
the files came in to late. While the author of the PostgreSQL solution
claims the wrong ZIP archive has been used.

Besides, was quite a simple database test, not too complicated queries,
very few stored procedures. No views, no triggers, no 2PC, no nothing.
So IMHO it was not even a good test of database extensibility.

So this comparison is really just... ah! forget about it! (Just please,
next time, let's put a little more effort into such a thing.)

Regards

Markus

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

Peter Eisentraut

unread,
Aug 28, 2006, 6:34:44 PM8/28/06
to
Scott Marlowe wrote:
> I'm wondering if the source code is available.

http://www.heise.de/ct/dbcontest/teilnehmer.shtml

> My guess is it was full of MySQLisms and the postgresql "port" was
> written without indexes, no transactions, and relied on running
> dozens of queries that postgresql could have more efficiently done in
> one query or in one transaction at least.

Your guess is wrong.

---------------------------(end of broadcast)---------------------------

Scott Marlowe

unread,
Aug 28, 2006, 6:34:52 PM8/28/06
to
On Mon, 2006-08-28 at 17:06, Markus Schiltknecht wrote:
> Tony Caduto wrote:
> > http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
> >
> > Don't know the validity of this dvd order test they did, but the article
> > claims Postgresql only did 120 OPM.
> > Seems a little fishy to me.
>
> Now, this article really s**ks! First of all, the original contest was
> specifically not only about performance. And the MySQL team did a whole
> lot of dirty tricks (i.e. using memcached) to push their solution.
>
> I am the one who has written he only PostgreSQL entry, for which I'm
> still sorry and ashamed, because it performs so poorly. I just didn't
> have much spare time to spend, but thought I'd send it in anyway. One of
> the reasons it did not perform well was, that I simply have forgotten to
> enable connection pooling.

Was this all the same basic task implemented by different teams then?

Can we see the code? hack it? I'm sure someone here could help out.

I don't care about the contest, but it would be nice to be able to put
out a version that could compete with MySQL's.

Scott Marlowe

unread,
Aug 28, 2006, 6:36:08 PM8/28/06
to
On Mon, 2006-08-28 at 17:34, Peter Eisentraut wrote:
> Scott Marlowe wrote:
> > I'm wondering if the source code is available.
>
> http://www.heise.de/ct/dbcontest/teilnehmer.shtml
>
> > My guess is it was full of MySQLisms and the postgresql "port" was
> > written without indexes, no transactions, and relied on running
> > dozens of queries that postgresql could have more efficiently done in
> > one query or in one transaction at least.
>
> Your guess is wrong.

Yes, I see that now. I thought it was one version of the same app
pointed at different DBs.

I see now that it's not. See my most recent post...

Markus Schiltknecht

unread,
Aug 29, 2006, 3:11:20 AM8/29/06
to
Scott Marlowe wrote:
> Was this all the same basic task implemented by different teams then?

Yep.

> Can we see the code? hack it? I'm sure someone here could help out.

Sure.

> I don't care about the contest, but it would be nice to be able to put
> out a version that could compete with MySQL's.

Sure. The main db-contest site of the magazine is [1], but it's all
german. You'lll find an english translation of the original article,
describing the problem to solve in [2].

The contest entries are downloadable on [3]. Mine is the 'bluegap' one.
If you want to hack together a high-performance solution based on
PostgreSQL I'd start with Alvar C.H. Freude's disqualified entry.

Happy hacking

Markus


[1]: http://www.heise.de/ct/dbcontest/
[2]: http://firebird.sourceforge.net/connect/ct-dbContest.html
[3]: http://www.heise.de/ct/dbcontest/teilnehmer.shtml

Robert Treat

unread,
Aug 30, 2006, 1:54:56 PM8/30/06
to
On Monday 28 August 2006 17:04, Peter Eisentraut wrote:
> Tony Caduto wrote:
> > http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
> >
> > Don't know the validity of this dvd order test they did, but the
> > article claims Postgresql only did 120 OPM.
>
> The contest evaluated the solutions sent in by whoever wanted to
> participate. This doesn't prove anything, except that the MySQL
> marketing department is exceptionally slow. These contest results were
> published months ago.

Seems they are fast where it counts... namely getting a team of developers on
top of the contest so they come out shining. Too bad people in this
community don't see the value of paying people to do advocacy related
development.

--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL

Martijn van Oosterhout

unread,
Aug 30, 2006, 4:11:37 PM8/30/06
to
On Wed, Aug 30, 2006 at 01:54:56PM -0400, Robert Treat wrote:
> Seems they are fast where it counts... namely getting a team of developers on
> top of the contest so they come out shining. Too bad people in this
> community don't see the value of paying people to do advocacy related
> development.

I don't remember this ever being mentioned on the lists? I never heard
of it. Maybe if it had been mentioned here it would've got more
attention?

It's true there isn't a group here looking for competitions to enter...

Have a ncie day,
--
Martijn van Oosterhout <kle...@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.

signature.asc

Joshua D. Drake

unread,
Aug 30, 2006, 4:48:15 PM8/30/06
to
Martijn van Oosterhout wrote:
> On Wed, Aug 30, 2006 at 01:54:56PM -0400, Robert Treat wrote:
>> Seems they are fast where it counts... namely getting a team of developers on
>> top of the contest so they come out shining. Too bad people in this
>> community don't see the value of paying people to do advocacy related
>> development.
>
> I don't remember this ever being mentioned on the lists? I never heard
> of it. Maybe if it had been mentioned here it would've got more
> attention?
>
> It's true there isn't a group here looking for competitions to enter...
>
> Have a ncie day,

I *think* Robert's point is that a paid advocate may have actually been
looking for this type of thing to come up and would help organize an effort.

Sincerely,

Joshua D. Drake

--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

---------------------------(end of broadcast)---------------------------

AgentM

unread,
Aug 30, 2006, 4:23:15 PM8/30/06
to

On Aug 30, 2006, at 16:11 , Martijn van Oosterhout wrote:

> On Wed, Aug 30, 2006 at 01:54:56PM -0400, Robert Treat wrote:
>> Seems they are fast where it counts... namely getting a team of
>> developers on
>> top of the contest so they come out shining. Too bad people in this
>> community don't see the value of paying people to do advocacy related
>> development.
>
> I don't remember this ever being mentioned on the lists? I never heard
> of it. Maybe if it had been mentioned here it would've got more
> attention?
>
> It's true there isn't a group here looking for competitions to
> enter...

Perhaps because PostgreSQL is too busy competing with itself? Is
anyone aware of any other open-source project that has added so many
new features in the past year?

-M

Robert Treat

unread,
Aug 30, 2006, 6:26:13 PM8/30/06
to
On Wednesday 30 August 2006 16:48, Joshua D. Drake wrote:
> Martijn van Oosterhout wrote:
> > On Wed, Aug 30, 2006 at 01:54:56PM -0400, Robert Treat wrote:
> >> Seems they are fast where it counts... namely getting a team of
> >> developers on top of the contest so they come out shining. Too bad
> >> people in this community don't see the value of paying people to do
> >> advocacy related development.
> >
> > I don't remember this ever being mentioned on the lists? I never heard
> > of it. Maybe if it had been mentioned here it would've got more
> > attention?
> >

FWIW it was mentioned and discussed on the german postgresql list.

> > It's true there isn't a group here looking for competitions to enter...
> >

> I *think* Robert's point is that a paid advocate may have actually been
> looking for this type of thing to come up and would help organize an
> effort.
>

Right... these types of things come up all the time, whether they be
benchmarks, contests, tutorials, or just general application support, and all
of the commercial database companies (ie. oracle, ibm, ms, mysql) have people
on staff to point at these tasks. The community has been good about full
time developers working on core, but seems to overlook all of the other areas
that you can point a paid developer that will increase adoption.

--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL

---------------------------(end of broadcast)---------------------------

Tim Allen

unread,
Aug 30, 2006, 10:27:28 PM8/30/06
to
Martijn van Oosterhout wrote:
> On Wed, Aug 30, 2006 at 01:54:56PM -0400, Robert Treat wrote:
>
>>Seems they are fast where it counts... namely getting a team of developers on
>>top of the contest so they come out shining. Too bad people in this
>>community don't see the value of paying people to do advocacy related
>>development.
>
>
> I don't remember this ever being mentioned on the lists? I never heard
> of it. Maybe if it had been mentioned here it would've got more
> attention?
>
> It's true there isn't a group here looking for competitions to enter...
>
> Have a ncie day,

Speaking for myself, I'm _delighted_ that the core postgres developers
spend their time improving postgres, rather than spending their time on
random beauty contests.

Tim

--
-----------------------------------------------
Tim Allen t...@proximity.com.au
Proximity Pty Ltd http://www.proximity.com.au/

0 new messages