Re: [HACKERS] \dt and disk access - Mailing list pgsql-hackers

From Vadim B. Mikheev
Subject Re: [HACKERS] \dt and disk access
Date
Msg-id 4eb240dd094385915ddd7773e1b80c06
Whole thread Raw
In response to [HACKERS] \dt and disk access  (Bruce Momjian <maillist@candle.pha.pa.us>)
List pgsql-hackers
Igor wrote:
>
> Well, in case you are wondering.....I populated a table with about 500
> tuples.
> Running select * on that table yealded no errors from purify..
> Running select * .. order by (a non-indexed field) gave the following:
>   total memory usage: 319,803 bytes.
> Of this amount of memory, 57,344 bytes were leaks.
                            ^^^^^^
57344/8192 = 7

>
> I then populated another table with 2600+ tuples.
> These are the results with order by:
>   Total memory used: 786,763
> Of this memory, 524,288 bytes are memory leaks!!! Waste of half a meg of
                  ^^^^^^^
524288/8192 = 64

I believe that it's just local buffers allocation, so it's not
real memory leak.

> ram on 2600 or so tuples is bad....And what if you have 20,000 tuples? Or
> a 100,000 ?

64 is max number of local buffers...

Vadim

------------------------------

pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: [HACKERS] Re: PostgreSQL psort() function performance (fwd)
Next
From: Bruce Momjian
Date:
Subject: Re: [HACKERS] Patch to psql.c to allow .psqlrc file