RANDOM RUMOURS (& TRUTHS?)
By S. Monger
Ah well, I’ve never claimed to be perfect, have I? In the last issue I
mentioned the Gemini GM886 board containing an iAPX186. Well it seems that
this idea was dropped a while ago, and in fact it is the GM888 that you will
see in coming months. What is the difference – well the GM886 board was to be
a board interfaced to the 80-BUS via a couple of I/O ports (like the GM812
IVC, GM832 SVC and IO828 Pluto). It contained its own 256K RAM and
consequently would have been very fast, and expensive! It would also have had
the disadvantage of not being able to drive any I/O peripherals directly, such
as Pluto, and would have to have passed on the task via the Z80 on the host
processor board, consequently slowing down the operation quite considerably.
After a little research Gemini discovered that they could produce a board that
would drive the 80-BUS directly, and thus be able to use all the I/O on the
bus. More importantly, from a cost viewpoint, the board would be able to have
NO memory on it, and would therefore almost certainly be under #200. Of
course in interfacing to the 80-BUS the processor used has had to be changed
to one with an external 8-bit interface, and so the 8MHz 8088 has been chosen.
So how do you use it? Well the system boots as normal under the Z80, you
then flip a bit of an 80-BUS I/O port, the 8088 puts out a bus request, the
Z80 finishes the instruction that it was executing, and then the 8088 takes
over control of the entire system. If at any stage you wish to pass control
back to the Z80 then the 8088 flips the I/O bit back, and the Z80 carries on
where it left off. In theory, therefore, you can write code that is of mixed
type, and switch to and fro between processors – must be of some use!
The GM888 board will also contain another of those blasted Real Time
Clock thingies, with battery back-up. This serves two purposes. First of all
certain 16 bit operating systems allow time and date stamping of files, and
secondly the actual chip chosen provides oodles of interrupts if required, and
this is necessary for task switching with Concurrent-DOS (previously called
Concurrent-CP/M) and presumably with Multi-tasking MS-DOS. There is also a
socket for the 8087 high speed arithmetic co-processor. Board availability?
And how about the extra memory that the 8088 can support? Well Gemini
have launched a 256K RAM board that supports the 80-BUS extended address
lines. It has several modes of operation:
|1)||4 pages of 64K from address 0 (for use with Nascom 2s and Gemini GM811s).|
|2)||4 pages of 64K from any extended address (EA).|
|3)||1 page (any) of 256K from any extended address.|
|4)||3 pages of 64K from 0 and 1 page of 64K from EA1 (allows GM813 users to
have 4 pages of 64K at 0 without contention.|
In addition to the above there is a ‘Common Area’ mode which puts a
common area of memory (selectable between 4K and 8K) in all four pages.
With all these modes it is possible to set up virtually any permutation
of page mode and extended addressing that any one may want, and systems with
the GM813 or GM888 will accept up to 8 boards = 2Mbytes! (4 pages of 512K.)
And the price of all this flexibility ? – #325 + VAT. Available now!!
With a distinct lack of signs of activity from Lucas, Gemini continues
its 80-BUS domination campaign! The Climax vector graphics colour board has
been taken over by Gemini, as the GM837. It is now only available in its
fully populated form (modulated and R-G-B outputs) and the price has been
reduced to #165 + VAT. Gemini has also taken over the IO Research 8-bit A-D
board, now the GM824, and the price remains the same at #125 + VAT. One 80-BUS
price has risen, the Belectra arithmetic board. It is now a totally
horrendous #268 + VAT!!