Re: Spinlocks, yet again: analysis and proposed patches

On Sun, 18 Sep 2005, Jim C. Nasby wrote:

> On Sat, Sep 17, 2005 at 01:40:28AM -0400, Tom Lane wrote:
> > Gavin Sherry <swm@linuxworld.com.au> writes:
> > > On Sat, 17 Sep 2005, Tom Lane wrote:
> > >> It'd be real interesting to see comparable numbers from some non-Linux
> > >> kernels, particularly commercial systems like Solaris.
> >
> > > Did you see the Solaris results I posted?
> >
> > Are you speaking of
> > http://archives.postgresql.org/pgsql-hackers/2005-09/msg00715.php
> > ?
> >
> > That doesn't seem directly relevant to the point, because it's for a
> > 2-CPU machine; so there's no way to run a test case that uses more than
> > one but less than all the processors.  In either the "one" or "all"
> > cases, performance ought to be pretty stable regardless of whether the
> > kernel understands about any processor asymmetries that may exist in
> > the hardware.  Not to mention that I don't know of any asymmetries in
> > a dual SPARC anyway.  We really need to test this on comparable
> > hardware, which I guess means we need Solaris/x86 on something with
> > hyperthreading or known NUMA asymmetry.
>
> I have access to a 4-way Opteron 852 running Solaris 10. What patches
> would you like me to test?

These ones here:

http://archives.postgresql.org/pgsql-hackers/2005-09/msg00566.php

Gavin