16 Jan
2003
16 Jan
'03
6:21 p.m.
Michael asks:
This last point makes me think "hmm, it must be
fairly generic, I
wonder what other VAXstations can it run on?"
True, the software set seems to
be fairly generic.. I found this out
a while ago by accident; I (net-)booted that image onto the wrong
machine, and dang! wasdat? it *runs* ??? :-)
Specifically I wonder whether it would run on a KA43
(VS3100 M76).
Most likely not, because the -M76 is an NVAX box similar to the 4000
series machines (MV4 aka 4000/200, the 4000's, the 3100-M76 and -4X
and -8X series.) It might run on the 3100-M38, although I believe
that's a "non" 3100, too. I would assume that they kept the CPU
support library as small as possible, meaning only the generic series
of machines they "sortof" intended the VXT software set for, i.e. the
2000, 3100 and alike systems and their hardware features (as needed.)
I'm willing to bet that it'll handle the SPX
video
Yup, it's basically some "NanoVMS" kernel, with minimal runtime
and a
VMS DECwindows subset. Which means (methinks..) that it most likely
wasnt stripped from its GPX and SPX(+) drivers.
Dang! Now you got me curious. I have all three (2000,M38 and M76) so
will set them up tonight or tomorrow and see what they do.
--fred
17 Jan
17 Jan
3:54 a.m.
On 2003.01.17 09:23 Fred N. van Kempen wrote:
Sorry, no. The m76 has a Rigel. It isn't even similar to the
VS4k. The
VS4k machines have different Ethernet (SGEC insted of LANCE) and SCSI
(asc insted of NCR).
VS4k VLC has SOC
VS4k60 has Mariah
VS4k90 and above have NVAX
MV4k200 has SOC
MV3k100m40 has SOC
MV3k100m8x has Mariah
VAX4k300 has Rigel
VAX4k400 and above as well as VAX4k1xx and MV3k100m9x have NVAX.
Taken from http://www.netbsd.org/Documentation/Hardware/Machines/DEC/vax/sections.html
Specifically I
wonder whether it would run on a KA43 (VS3100 M76).
Most likely not, because the
-M76 is an NVAX box similar to the 4000
series machines It might run on the 3100-M38, although I believe
that's a "non" 3100, too.
VS3100 m30/38/40/48 are almost identical.
The only differences are the
enclosure and speed.
I would assume that they kept the CPU
support library as small as possible, meaning only the generic series
of machines they "sortof" intended the VXT software set for, i.e. the
2000, 3100 and alike systems and their hardware features (as needed.)
Looks so. It
seams to support only NMOS-VAX (VS2k), CVAX (VS3100
m30/38/40/48) and maybe SOC. I think SOC is the CPU in the VXT2k.
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
4:41 a.m.
MV3k100m8x has Mariah
Just to nitpick but although the UV3100-80 is indeed
based on the Mariah chipset, the UV3100-85 and
UV3100-88 are both NVAX-based IIRC. The -85 has the
NVAX VIC disabled by the console during initialisation.
I forget whether the -88 was slugged in the same way or not.
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
6:25 a.m.
On 2003.01.17 19:40 Antonio Carlini wrote:
Just to nitpick but although the UV3100-80 is indeed
based on the Mariah chipset, the UV3100-85 and
UV3100-88 are both NVAX-based IIRC.
Ahh, thanks for the correction. I just
extrapolated.
The -85 has the
NVAX VIC disabled by the console during initialisation.
VIC???
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
7:18 a.m.
FWIW my recollection is that the UV3100-88, UV3100-98
and VAX 4000-108A are the same hardware as the UV3100-85,
UV3100-96 and VAX 4000-106A but respun to fit in a
PC-style case (and, for at least some options, using
PCI connectors but *NOT* a PCI bus).
Virtual Instruction Cache. The NVAX cache hierarchy goes something
like: 2KB VIC, an 8KB P-cache (on-chip: can do Instructions, Data
or both according to config) and up to 2MB of write-back B-cache
(this one is external to the NVAX chip itself).
It's possible to enable and disable the various caches. The VIC can
be disabled by "patching" the NVAX in some dynamic way (presumably
non-reversibly or at least only reversible in an obscure way). I
don't think the same can be done for the P-cache an the B-cache is
physically fitted to the board so, short of grabbing a soldering iron,
you are stuck with what the system came with.
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
The -85 has
the
NVAX VIC disabled by the console during initialisation.
VIC???
9:14 a.m.
On 2003.01.17 22:20 Antonio Carlini wrote:
FWIW my recollection is that the UV3100-88, UV3100-98
and VAX 4000-108A are the same hardware as the UV3100-85,
UV3100-96 and VAX 4000-106A but respun to fit in a
PC-style case (and, for at least some options, using
PCI connectors but *NOT* a PCI bus).
And the MV3100m9{0,5,6} are identical to the
VAX4000m10{0,5,6}. The
difference is that the MV models can't use the QBus adapter (on the base
board) and have no DSSI (on an opton board?).
Sad that I have not the QBus cabling for my MV3100m95. A 83 MHz NVAX (32
VUPs?) desktop VAX wirh QBus would be nice.
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
9:36 a.m.
And the MV3100m9{0,5,6} are identical to the
VAX4000m10{0,5,6}. The difference is that the MV models can't
use the QBus adapter (on the base
board) and have no DSSI (on an opton board?).
The mainboards are the same (i.e. a KA50 *is* a KA52 etc.)
and there is a console test that switches the identity
back and forth. So if you switch the identity you can use the
Qbus interface and the DSSI. Obviously you have to have a
Qbus extension cable and some way of connecting it - the easiest
way is a VAX 4000-1xx cab :-) I don't remember whether
DSSI needs an additional board which is part of the VAX 4000-1xx
cab or whether it too just works.
Sad that I have not the QBus cabling for my MV3100m95.
A 83
MHz NVAX (32
VUPs?) desktop VAX wirh QBus would be nice.
It would indeed. The VAX 4000-1xx range is something that
I only rarely got to play on. But there are plenty of
other Qbus boxes around.
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
18 Jan
18 Jan
2:08 a.m.
On Fri, Jan 17, 2003 at 11:38:58PM -0000, Antonio Carlini wrote:
I have a MV II, MV III, MV3900, MV4k200, VAX4k300,
VAX4k400
and a PDP11-73. Enough QBus boxen to play. OK. I miss a KA640
based machine. But none of them brings that power in that small
footprint and is easy to interface to lots of storage, i.e.
SCSI.
--
tschüß,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz
Obviously you have to have a Qbus extension cable and
some way
of connecting it - the easiest way is a VAX 4000-1xx cab :-)
Well, the enclosure
is the same. It is only a set of additional
connectors on the back an internal cabling. At least I interpreted
pictures of a VAX4k105 this way. But all I need is the pinout.
Building some cables is no problem and I have a QBus backplane
from a dead and rotten BA23.
Sad that I
have not the QBus cabling for my MV3100m95. A 83
MHz NVAX (32 VUPs?) desktop VAX wirh QBus would be nice.
It would indeed. The VAX
4000-1xx range is something that
I only rarely got to play on. But there are plenty of
other Qbus boxes around.
6:48 a.m.
Well, the enclosure is the same. It is only a set of
additional
connectors on the back an internal cabling. At least I interpreted
pictures of a VAX4k105 this way. But all I need is the pinout.
Building some cables is no problem and I have a QBus backplane
from a dead and rotten BA23.
I have a UV3100-90 and I've been close to a VAX 4000-100 (it
just wasn't mine). The VAX 4000-100 enclosure has a distinctive
sloping part at the back which houses the Qbus (and DSSI?)
connectors. I didn't get to look closely enough to see whether
this housing is an add-on to the UV3100-90 enclosure or the whole
case is a manufacturing modification of that enclosure. Either way,
the standard shipping UV3100-90 enclosure does not have anywhere
for Qbus and DSSI connectors.
I'll look for pinouts if I can dig up the right docs.
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
8013
days inactive
8014
days old
8 comments
3 participants
participants (3)
-
arcarlini＠iee.org -
Fred.van.Kempen＠microwalt.nl -
jkunz＠unixag-kl.fh-kl.de