23 Nov
2015
23 Nov
'15
7:18 p.m.
Perhaps Oberon required something faster than the 70ns - 85ns tRC read
cycle / tWC write cycle times supported by the Micron cellular RAM
running in asynchronous mode on the Nexys 2 - presumably because the
data path is only 16 bits?
That said, most "retro" computing would, I expect, be perfectly happy
with that kind of cycle time and bus width, yes? And the implementation
in this asynch. mode is very very easy - classic SRAM interface signals.
Looks like the Nexys 3 and Nexys 4 boards have cellular RAM like this as
well.
JRJ
On 11/23/2015 8:47 PM, William Maddox wrote:
[Yahoo's webmail client garbled the last link --
resending]
The revived 2013 re-issue of Niklaus Wirth's Oberon system is a joy to behold. If
you've never heard of Oberon before, it is a minimalistic education-oriented language
and operating system designed after Wirth had taken a (second) sabattical at PARC in the
80's.
The new version runs on a custom RISC processor, implemented in an FPGA, instead of the
NS3032 in the orginal Ceres workstations. Originally, it required a Digilent
"Spartan 3 Starter Kit" with a custom-built daughterboard providing a few
additional connectors. This board is no longer made, however, and no other FPGA
development board appears to provide the 32-bit wide fast SRAM the Oberon CPU required.
Recently, a new board, the OberonStation, has come onto the market that was designed
specifically for Oberon, and will boot up Oberon 2013 out of the box. It also looks like
an excellent platform for other retro-style FPGA CPU designs that want to stay away from
complex SDRAM controllers and the caches they like to feed.
My OberonStation arrived a couple of days ago, and it's really amazing to see what
can be done with a hardware and software stack that is small enough to actually read and
understand.
https://www.inf.ethz.ch/personal/wirth/
http://www.projectoberon.com/
http://oberonstation.x10.mx/