7 Aug
2001
7 Aug
'01
1:14 p.m.
At 07:32 PM 8/7/01 +0100, Tony wrote:
FWIW, FPGA-based re-implementations of classic
computers don't satisfy my
interests either. For 2 main rasons :
1) I don't have any machine (read : lusedoze box) that will run the
development tools, so I can't get in there and 'tinker'
You assume too much Tony. The tools run on UNIX systems as well as Windows
machines. If you read the page on the PDP-8/X you will note that the author
wrote _his own_ tools and those of course you could run under RSX-11M.
Further it would be possible to port them to Fortran but that might be too
painful.
2) You can't clip a logic analyser onto the
internal signals of an
FPGA... Remember I think of computers in terms of gates and flip-flops...
Actually this isn't quite true either. Especially with respect to the
Xilinx RAM based FPGAs. Those nets on your schematic can be "bound out" to
any spare I/O pin. Generally I/O isn't a killer if you're using a BGA package.
I am not at all convinced that there will be a time
when it will be
impossible to keep something like a PDP11 running, either. All the chips
in something like an 11/10 or 11/45 are documented. You could, if you had
to, re-implement a single chip in an FPGA. Sure it would be a waste of
FPGA, but if the time came when you couldn't get (say) a 74S181, and you
needed one, that would be a way of keeping the machine operational.
There is the concern however that the only FPGAs you could get would run at
3.3V or worse 1.8V. The smallest one would probably be a 36pin SOIC package
that you would have to dig out your old microscope to solder on to an
adapter board of some type. I've suggested for a while that the Xilinx
folks should take their 9500 CPLD series, put it into 14, 16, 18, and 22
pin packages, surround the die with 74S TTL compatible i/os and then have a
machine that allowed you to type in the TTL part you wanted and push the
button. And "poof" you'd get one of those. They don't think there would
be
enough of a market to justify the tooling costs.
--Chuck