[comp.sys.apollo] Apollo DN3000

krowitz@RICHTER.MIT.EDU (David Krowitz) (02/04/91)

A few minor corrections ...

The DN3000 is a 12 MHz 68020/68881 CPU running at about 1.5 MIPS and 0.07 MFLOPS.
It is not an 68030 CPU.

The MMU unit depends on the version of the CPU. Earlier versions of the DN3000
were produced before Motorola had their PMMU chip (the 68551?) available. These
machines were limitted to a 64MB virtual address space per process. Under SR10.3
you may have a maximum of 128 processes on one of these early models. Later models
(the DN3010 and DN3010A) have the Motorola PMMU chip. These machines support up to
256 processes with larger address spaces (at least 256MB/process). Of course on
all machines, you are limitted by the amount of disk space available for VM.

You do not necessarily need a co-processor board to run DOS application. DPCC
(DPCE?) is a DOS emulator which will run most DOS applications. However, running
a 80x86 emulator on a 12 MHz 68020 will be much slower than a 80286 or 80386
co-processor card (obviously).

The disk drives are ESDI 5 1/4" full height drives which come in 86 MB, 155 MB, and
348 MB capacities. A full SR9.7 OS will leave about 30 MB free on a 155 MB drive.
A full Aegis plus BSD SR10.2 OS will leave about 5 MB free on a 155 MB drive. SR10.3
can not be loaded on a 155 MB drive in a full configuration (neither Aegis-only, BSD-only,
nor SYSV-only will load from tape without a hand-tailored install). The bright side
to this is that once you've loaded the OS onto a single machine, most of the other
machines in the network can have much of their OS soft-linked to the first machine
(ie. man pages, debugging tools, additional fonts ... anything not critical to the
booting of the machine). You can configure a local disk with as little as 40 MB of
the OS. This is not very impressive compared to DOS, but it's a lot better than most
Unix systems.

Keep in mind that the DN3000 was first introduced in the Spring on 1985. The machine
is now almost 6 years old, which is about 3 years beyond the expected lifetime of
a product in the technical workstation market. (ie. Ahyee! another typo ... that should
be 1986 and 5 years! ... I *got* to stop running backups while typing mail responses!).



 -- David Krowitz

krowitz@richter.mit.edu   (18.83.0.109)
krowitz%richter.mit.edu@eddie.mit.edu
krowitz%richter.mit.edu@mitvma.bitnet
(in order of decreasing preference)