|
Authored by: JamesK on Monday, April 15 2013 @ 05:31 PM EDT |
{
Add to the fact that CP/M was designed to cope with machines with varying
amounts of memory. It wouldn't have suffered from the 640Kb problem ...
}
I could be mistaken (it has happened before), but IIRC that problem was created
by IBMs design and not inherent in DOS. Of course the 8086 & 8088 were
limited to 1 Mb anyway and there wasn't any support for memory management beyond
that, until after market memory expansion boards appeared. As for CP/M-80, it
was limited to 64K, but there were some memory boards that supported bank
switching available. I experienced similar with the Data General computers.
They could handle up to 32K words or 64K bytes, but then a memory mapping board
was developed that could be used to greatly expand the available memory, moving
it around in (IIRC) 2K word chunks, though the CPU only saw 32K at at time.
---
The following program contains immature subject matter.
Viewer discretion is advised.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, April 15 2013 @ 06:36 PM EDT |
> It wouldn't have suffered from the 640Kb problem ...
CP/M was for 8080 type CPUs and these were limited to accessing 64Kb. MP/M (in
1978) was also 8080 but could do bank switching with an external memory
switching scheme (ie the CPU only saw 64Kb - usually 16Kb fixed plus 48Kb
switchable. I have a machine or two around here that could use 256Kb for
multi-user/multitasking plus 256Kb RAM Drive.
CP/M-86 was for 8086/8088 which had an address limit of 1Mbyte. The 640Kb was
not the result of MS-DOS but was a limit of the IBM PC architecture which put
memory mapped devices above 640Kb, for example the Mono and CGA screen buffers.
MS-DOS could also use almost 1Mbyte on suitable machines, such as S100 bus
systems.
Concurrent-CP/M-86 on 8086/8088 machines could use bank switching to access
several Mbytes with EEMS cards. Some machines, such as Apricot had EEMS built
in. I used AST RAMPage cards and Arnet 8port serial cards to get multiuser on an
IBM-PC XT (as well as other maker's machines).
Later Concurrent-DOS and Multiuser-DOS ran on 386 and better which solved the
memory issue entirely by running multiple 8086 mode sessions for the programs.
[ Reply to This | Parent | # ]
|
|
|
|
|