...
The Meta 4 (from Digital Scientific) was a clone of the
IBM 1130, made in San Diego (ok, maybe Del Mar or Sorrento
In fact, the Computerwoche article above maintains it is San Diego. I've
seen San Diego in other Meta 4 references I've found like the one Stan
offers below.
The headquarters and factory were in the Sorrento Valley, although their
mailing address may have been San Diego. I visited their offices once. I
supported a DSC Meta-4 for several years, about 1973 to 1976. It replaced
an IBM 1130 which it emulated exactly, but was 10 times faster when running
IBM 1130 FORTRAN IV programs. It was an elegantly simple machine. The
micro controller was truly a RISC machine, I think it had a 3 bit op code, 8
instructions.
>Valley area) around 1970. IIRC it was called an
1130 clone, but
>actually had the extra instruction(s) that would really make it an 1800
>clone.
Yes it had the set/clear memory protection bits. The memory arrays in
IBM
1130/1800's were 18 bits ... 16 data bits and 2 parity bits for an 1130 and
16 bits, one parity bit and one memory protect bit for the 1800.
>
>I remember that the Meta 4 had firmware that was implemented on
>boards about 1 foot by 1 foot, with little copper squares of foil about
>the 1/4" by 1/4" ...indicating 1/0 by presence/absence. One problem
>was that the squares would sometimes lift up a bit, so we'd take out
>the boards and press them down again.
Yes it was a type of capacitor Read Only Storage with a read cycle of 90 ns.
I remember we
also had APL, on a removable disk cartridge.
In the pages I've seen so far, the 1130 is said to have a max of 16KWord of
memory but my professors at school were bragging about our machine having
32K of memory (in 1972). Was there in fact an upgrade to 32K _words_ or
were they simply getting 32K bytes and 16K words confoozed? Remember, this
was all quite new to these older age professors then when few
backwater-area colleges our size even had a computer.
The 1130/1800 addressed 32K words, but the DSC Meta-4 had a feature so that
you could load 64K or RAM, then context switch between two 32K banks. In
this manner they developed a multi user 1130 timesharing system. The
timesharing control system was in the upper 32K and it swapped out 1130
users. If you had a fixed head disk attached, a feature available only on
the DSC Meta-4, not the 1130, you could swap out the whole 32K in some
ridiculously short time, like 50 ms., then swap in another 32K user in 50
ms. and pick up where they left off. It really worked, we attached 8 dumb
terminals to it and had 8 simultaneous 1130's running for less than the
price of a single IBM 1130, and since it was 10 times faster than an 1130,
each user could be running at about the same speed as an 1130, if they were
compute bound ... not usually because most of the users were pounding the
keyboard editing FORTRAN code and compiling. It worked well until we
replaced it with a DEC SYSTEM-20 which was 10 to 100 times faster still ...
at about the same price.
I still have the manuals for the Meta-4, even though I was dumb enough to
get rid of all of my 1130 manuals.
As an aside, DSC developed the line of National Semiconductor IBM Mainframe
clones that are now one of the Japanese lines ... and disappeared.
-- Dean
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Dean Billing Phone: 530-752-5956
UC Davis FAX: 530-752-6363
IT-CR EMAIL: drbilling(a)ucdavis.edu
One Shields Way
Davis, CA 95616