On Sat, 26 Apr 2003 18:30:30 +0100
Chris Gilbert <chris@dokein.co.uk> wrote:

> (note replying to port-arm only with this)
> 
> On Thu, 17 Apr 2003 08:05:44 +0100 (BST)
> Steve Woodford <scw@wasabisystems.com> wrote:
> 
> > On Thu, 16 Apr 2003 cgd@broadcom.com wrote:
> > 
> > > At Wed, 16 Apr 2003 23:27:28 +0000 (UTC), "Steve Woodford" wrote:
> > > > So, what are the main features/benefits of the new pmap?
> > >
> > > Have you benchmarked the actual impact on performance?  If so,
> > > what are the results?
> > >
> > > (If not, well, how do you know that things are really faster?  8-)
> > 
> > Heh, as Jason said, the new pmap has been benchmarked on a 600MHz
> > i80321 xscale cpu. Here is a quick summary:
> 
> A quick whirl on my cats (with my favorite test: make cleandir; make
> buildlink; time make configure for pkgsrc/devel/gmake)
> gives the following times:
> old pmap:
>       176.56 real        74.60 user       120.59 sys
> new pmap:
>       149.15 real        75.67 user        93.10 sys
> 
> So it's most definetly faster.   I'll try and dust off my acorn32 and
> do the needed work, and get figures together for that.

on my acorn32 with a GENERIC kernel (no optimisation for SA110 etc)
old pmap:
	223.44 real        59.98 user       183.52 sys
new pmap:
	204.76 real        60.35 user       162.09 sys

SA110 optimised, with NEW_BUFQ_STRATEGY, and soft dep mounted disks:
	186.33 real        63.59 user       141.78 sys

Note that the userland on the box dates back to august 2002, so I maybe
missing the latest and greatest speeds ups from ld.elf_so as well.

Chris