--pgp-sign-Multipart_Mon_Oct_30_19:16:57_2006-1
Content-Type: text/plain; charset=US-ASCII

>>>>> "sc" == Sean Caron <caron.sean@gmail.com> writes:
>>>>> "jk" == Jochen Kunz <jkunz@unixag-kl.fh-kl.de> writes:

    sc> toss that sparc in the corner and it'll run for years and
    sc> years and years

yeah, i agree.  and for now I still use them.  even if we can fix
thread bugs and PCI bugs, it just seems like the downsides won't be
going away, with many of them like power consumption, bugs/regressions
in other free software, discriminatory proprietary software pricing
practices, ``blob'' drivers for hardware (the blob is in the BIOS
chip), not even under our control.

    sc> i don't know what i'm gonna do when my current line of
    sc> equipment goes south. i don't think netbsd is ever gonna
    sc> support the ultrasparc III.

I heard at NYCBSDCon an OpenBSD guy gave a talk on their more-or-less
mostly-working UltraSPARC III support?  but...

some have things like ECC on the L2 cache.  That's sort of an very
desireable toss-in-the-corner feature, because it means you can find
and fix a problem really fast rather than futzing around with it like
a gamer/overclocker.  as if they were meant for sale like mainframes
with an uptime SLA or something.  ...though, the opteron stuff
advertises L2 cache parity now, too, and it seems like some of these
temperature/ecc/parity things that print out ``DIMM 3 detected
correctable error'' on OSF/1 or Solaris will just print ``machine
check 1234 ip=0xffffffff234abcd Halting system'' on most free OS's.

    jk> I use LFS (Log-structured File System) on my U10 with great
    jk> success.

on sparc64?  that's good.  for how long have you used it without
kernel panics and/or fsck_lfs processes that consume unreasonable
amounts of memory and then die with no error message?  Do you use it
for /usr and homedirs, or just as an objdir scratch space for doing
builds of NetBSD over and over?  That's pretty significant if it works
well now.  Most of the success I've had myself and heard of until now
is the latter---it gets tried out for builds and never graduates to
real use.

--pgp-sign-Multipart_Mon_Oct_30_19:16:57_2006-1
Content-Type: application/pgp-signature
Content-Transfer-Encoding: 7bit

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

iQCVAwUARUaWAonCBbTaW/4dAQIf1AQArDDRxqzBsl9k3EjvFL3Gl7dKmfnISLdw
7+D+f0rOjX1xY2D1prlNJsgRXmqic5akZTTZ1UGp8XMHscZAFc/k3sD8fHflurUy
xfQldANHbXkvfdVyLT5oshGblX3UnRvJQL+YL+GZjTZidIdjx3PBI4NhYtRWfLp7
wloMp4ED5aQ=
=QFjI
-----END PGP SIGNATURE-----

--pgp-sign-Multipart_Mon_Oct_30_19:16:57_2006-1--