Обсуждение: Jav Database Performance

Поиск
Список
Период
Сортировка

Jav Database Performance

От
"Walter Cruz"
Дата:

Re: Jav Database Performance

От
Markus Schiltknecht
Дата:
Hi,

Walter Cruz wrote:
> Have anyone seen this?
>
> http://devloop.org.uk/documentation/database-performance/

Looks like a good, well documented benchmark. Was well worth the read.
Probably even more if you're more into Java than I am.

What's surprising me is, that PostgreSQL often was 'only marginally
slower' compared to MySQL. First, I thought: wow, MySQL must have
improved. But later on, I've read that they tested the MyISAM storage
engine. Thus, I'm surprised that

  a) they perform only marginally better, and

  b) that the results are so close and similar.

I can only conclude, that PostgreSQL does a very well job in keeping the
costs of consistency very low. ;-)

Regards

Markus

Re: Jav Database Performance

От
"Joshua D. Drake"
Дата:
Markus Schiltknecht wrote:
> Hi,
>
> Walter Cruz wrote:
>> Have anyone seen this?
>>
>> http://devloop.org.uk/documentation/database-performance/
>
> Looks like a good, well documented benchmark. Was well worth the read.
> Probably even more if you're more into Java than I am.
>
> What's surprising me is, that PostgreSQL often was 'only marginally
> slower' compared to MySQL. First, I thought: wow, MySQL must have
> improved. But later on, I've read that they tested the MyISAM storage
> engine. Thus, I'm surprised that
>
>  a) they perform only marginally better, and
>
>  b) that the results are so close and similar.
>
> I can only conclude, that PostgreSQL does a very well job in keeping the
> costs of consistency very low. ;-)

Or:

We Rock!

Joshua D. Drake


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


--

      === 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/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/


Re: Jav Database Performance

От
Florian Weimer
Дата:
* Markus Schiltknecht:

> I can only conclude, that PostgreSQL does a very well job in keeping
> the costs of consistency very low. ;-)

And the cost of indexes.  Which makes me doubt whether there is a
systematic error in the benchmark.  Perhaps this is a result of the
small working set.

Re: Jav Database Performance

От
Markus Schiltknecht
Дата:
Hi,

Florian Weimer wrote:
> And the cost of indexes.  Which makes me doubt whether there is a
> systematic error in the benchmark.  Perhaps this is a result of the
> small working set.

Uh.. I thought they probably have the same indexes on MySQL as on
PostgreSQL...

Markus

Re: Jav Database Performance

От
Florian Weimer
Дата:
* Markus Schiltknecht:

>> And the cost of indexes.  Which makes me doubt whether there is a
>> systematic error in the benchmark.  Perhaps this is a result of the
>> small working set.
>
> Uh.. I thought they probably have the same indexes on MySQL as on
> PostgreSQL...

No, I meant the comment on the bottom of page 11.  But I'm not sure if
the data backs it, the diagram is so hard to read.

Re: Jav Database Performance

От
Josh Berkus
Дата:
Walter,

> http://devloop.org.uk/documentation/database-performance/

Hmmm ... we could use someone to run a properly configured PostgreSQL on his
test. I'm pleasantly surprised by how well we did against other DBs with *no
performance tuning*.  Cool!

--
Josh Berkus
PostgreSQL @ Sun
San Francisco

Re: Jav Database Performance

От
Jim Nasby
Дата:
On Apr 23, 2007, at 8:36 PM, Josh Berkus wrote:
>> http://devloop.org.uk/documentation/database-performance/
>
> Hmmm ... we could use someone to run a properly configured
> PostgreSQL on his
> test. I'm pleasantly surprised by how well we did against other DBs
> with *no
> performance tuning*.  Cool!

Better yet would be if we shipped with different configs so they
could at least pick a more reasonable configuration for their hardware.

Something else that comes to mind... if they're running the default
config, they're not running autovacuum; so were they vacuuming at all?
--
Jim Nasby                                            jim@nasby.net
EnterpriseDB      http://enterprisedb.com      512.569.9461 (cell)