--TmwHKJoIRFM7Mu/A
Content-Type: text/plain; charset=iso-8859-15
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Feb 15 2006 - 06:53, Ben Collver wrote:
> Hello,
>=20
> One of the reasons I got sparc64 going was to "stress-test" my software
> and check for assumptions about memory alignment and number formats.
> The bus error you got likely indicates such an assumption.  From what I
> have read, you may see a performance improvement if you use the 32 bit
> sparc userland and tools with a 64 bit kernel.  I've been told to expect
> problems, ie: NetBSD threads won't work.
>=20

Don't know why but I had in my mind that running netbsd/sparc binaries
on netbsd/sparc64 was not possible... I have run netbsd/i386 binaries
=66rom a netbsd/amd64 installation but thought it would not work on
sparc/sparc64.

Because running the few things that cores when compiled sparc64 with
sparc emulation would be OK for me... If that works :)

> > LDAP or HTTPS server. From what I saw, using /i386 and /amd64 on my
> > Athlon64 3200+ with 512Mo of RAM, you don't get improvement worth the
> > lack of support for 64bits (soft that don't existe in 64bits like flash
> > plugin or things that cores like firefox).
>=20
> I have never had firefox coredump on amd64.  When I ran benchmarks, I

I meant on sparc64. Firefox on amd64 is OK.

> found a rough performance loss of 15% when running the i386 port on
> amd64 hardware.

--=20
,- This mail runs ------.
`--------- NetBSD/smtp -'

--TmwHKJoIRFM7Mu/A
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (NetBSD)

iD8DBQFD80WL0/VH7L7F7Y4RAlzuAKCJdK87vLGbJnyTT1U9a//6N6eenwCfeApT
id7APKdKKBmD6LfqZIMUSbc=
=WGjl
-----END PGP SIGNATURE-----

--TmwHKJoIRFM7Mu/A--