> -----Original Message-----
> From: Eran Tromer [mailto:[EMAIL PROTECTED] 
> Sent: Friday, April 29, 2005 11:46 AM
> To: Tzahi Fadida
> Cc: 'Linux-IL mailing list'
> Subject: Re: cleaning memory.
> 
> 
> On 29/04/05 05:52, Tzahi Fadida wrote:
> 
> > Postgresql only uses regular files as its db.
> 
> Then my first suggestion is inapplicable.
> 
> 
> > Sadly postgresql relies on the underlying OS caching. 
> Readaheads and 
> > all
> 
> But aren't you trying to (selectively) disable OS caching?

Its just so I can see the bottlenecks and do the experiments.
On the large scale when the algorithm is run with other queries,
the cache would be depleted anyway in favour of memory allocations
and compared to the size of the input it would be useless because
of sequential scans.
The thing is, for the experiments I can't run the algorithm on 200mb
relations
(because I have 192mb memory)
since it would finish in the next century. and I need to run it hundreds
of times.
I was thinking of just locking away memory but it is very tricky to lock
away
just the right amount of memory so the cache would be small enough to
be beaten by sequential scans of lets say 2-3mb relations.
I'll have to run test to see if I c reads when sequentially accessing
the algorithms
each time. In addition I'll have to disable all services so the memory
won't fluctuate
too much.

My last resort is to try to change the code of postgresql to use the
O_DIRECT but
its never so simple as just changing an option when opening in one
place.
>From the code I see they have checks in other places and there are
adjustments
to accommodate portings, etc...

> 
> 
>   Eran
> 
> 



=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to