14 May
2012
14 May
'12
9:55 a.m.
On May 14, 2012, at 12:26 PM, Chuck Guzis wrote:
On 14 May 2012 at 17:05, Andriy Romanenko wrote:
Since we're talking VHDL here, how does the performance of the old
USSR chips differ from the performance of a modern FPGA
implementation, say, the W11A version of the PDP 11/70 on opencores?
I don't know about specific implementations, but I do know FPGAs pretty
well. I'd feel pretty safe saying you should be able to get them to
the 10s of MHz fairly easily, even on low-cost FPGAs, if you do your
homework right. If you're trying to duplicate the internal workings
of the older chips (i.e. to pass CPU-specific diagnostics), it might
wind up a bit slower because you'll have to force some design trade-
offs. Among other things, there are no real tristate buffers inside
the FPGA, something which a lot of processors rely on for internal
buses.
I'm pretty sure the PDP2011 project winds up pretty close to the real
chips, even on low-end FPGAs.
- Dave
...have
three of them and some VM2 too. I where interested in such a
project. Currently I'm trying to build ans SBC out of an VM2 with
some old Tesla MVB1012 Chips (AY5-1013?) and a CPLD...
could you share some schematics/soft/vhdl you have done for this
project?