spike1@xxxxxxxxxxxxx wrote:
> Peter Köhlmann <peter.koehlmann@xxxxxxxxxxx> did eloquently scribble:
>>> I bet it doesn't take long for the emulator to appear. :)
>
>> You bet wrong. The hardware would not be easy to emulate, and it would
>> make the system slow to boot
>
> Ah gowan
> If you can emulate an x86 or a motorola 680x0, how much more difficult
> would it be to emulate one of those mainframes?
>
> FROM THE 1960s?
Actually, the first MULTICS version was available in 1969, so it wasn't
exactly just "from the 60s"
Additionally, the mainframes gained quite some speed too until the end of
MULTICS
> OK, the only difficulty in writing an emulator to run them on is reference
> hardware isn't available to reverse engineer.
The problem is that that hardware mainly was to secure the processes (and
users) from the others.
That would be difficult to implement as emulator, since to *really* emulate
it you would have to write wrappers around all memory accesses as well
> I wouldn't be at all surprised if a 1986 Sinclair ZX Spectrum 128 would've
> given one of those mainframes from the '60s a run for their money.
You would be surprised. They *still* run rings around PC hardware
The I/O performance those machines had is still awesome
A *single* process, CPU bound, might run faster now on a modern x86
processor. But there are no single CPU bound processes when you consider
speed of a *system*
--
Windows: Because everyone needs a good laugh!
|
|
