Hi Joe,
the general rule on Solaris SPARC is:
- if you need to address a big size of memory (over 4G): compile in 64bit
- otherwise: compile in 32bit ;-)
It's true that 32bit code will run faster comparing to 64bit ont the
64bit SPARC - you'll operate with 2 times shorter addresses, and in
some cases SPARC will be able to execute 2 operations in parallel on
32bit code, while it'll be still one operation on 64bit code.. - But
it's all about the code, because once you start to do I/O requests all
kind of optimization on the instructions will be lost due I/O latency
;-))
So, as usual, a real answer in each case may be obtained only by a real test..
Just test both versions and you'll see yourself what is a valid in
your case :-))
Same problem regarding compilers: in some cases GCC4 will give a
better result, in some cases Sun Studio will be better (there are many
posts in blogs about optimal compiler options to use).. - don't
hesitate to try and don't forget to share here with others :-))
Rgds,
-Dimitri
On 8/11/10, Joseph Conway <mail@joeconway.com> wrote:
> With a 16 CPU, 32 GB Solaris Sparc server, is there any conceivable
> reason to use a 32 bit build rather than a 64 bit build? Apparently the
> Sun PostgreSQL package includes a README that indicates you might want
> to think twice about using 64 bit because it is slower -- this seems
> like outdated advice, but I was looking for confirmation one way or the
> other.
>
> Also semi-related question: when building from source, using gcc,
> enabling debug (but *not* assert) is normally not much of a performance
> hit. Is the same true if you build with the Sun CC?
>
> Thanks in advance for any thoughts/experiences.
>
> Joe
>
>
>
> --
> Joe Conway
> credativ LLC: http://www.credativ.us
> Linux, PostgreSQL, and general Open Source
> Training, Service, Consulting, & 24x7 Support
>
>
> --
> Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-performance
>