Re: #entries in APSW statement cache

Ed Pasma <[email protected]>
Newsgroups gmane.comp.python.db.pysqlite.user
Message-ID <[email protected]>
Hello Roger,

The test where the size of the statement cache matters so much is  
actually a real application, although a hobby-one. It computes chess  
moves, purely SQL based with some procedural control from Python. It  
is hard to send this. The effect of increasing the statement cache is  
that the elapsed time is 35 % less. This is quite particular but we  
don't know what normal use is and this is just a case. DRH is right  
that SQLite parses incredibly fast, I estimate a factor 100 compared  
to Oracle. But I would not be surprised if he has stated this very  
careful with room for extreme cases. By the way it is faster than  
pysqlite now :)

Regards, Ed








Op 7-dec-2007, om 21:42 heeft Roger Binns het volgende geschreven:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Ed Pasma wrote:
>> In a test where the speed is compared,
>
> Could you post or mail me the speed test?  I always like that sort of
> code since it helps with profiling etc.
>
>> match very well. It appears that this is due to the fact that the
>> test includes more different SQL statements than what fits in the
>> APSW statement cache.
>
> The statement cache should end up with the most used statements.  Do
> real world programs really have that many different "hot" statements?
> DRH has also claimed that SQLite is really fast at compiling  
> statements.
>  That is why I sized the statement cache at 32.
>
>> Could this be increased to 100 as in pysqlite?
>
> Done.
>
>> Or could you make the number of entries in the cache configurable?
>
> Yes, I have put in a ticket for this and will do it before the next
> release.  (The only thing remaining for the next release is ripping  
> out
> all the stuff to do with threading).
>
> Roger
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFHWbAtmOOfHg372QQRAng8AJ9GnCivLjG6VEKpIAtLbrWofR3lyQCffVzG
> 44ACfPuh8loY80fTlOisO2g=
> =mpVp
> -----END PGP SIGNATURE-----
> _______________________________________________
> pysqlite mailing list
> pysqlite-IAPFreCvJWPBWskQ1e/[email protected]
> http://lists.initd.org/mailman/listinfo/pysqlite
>
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.