[Oberon] Cheap chinese FPGA board - is it usable for PO?
Paul Reed
paulreed at paddedcell.com
Mon Nov 11 18:33:39 CET 2019
> I do not even know whether or not adding a cache would affect the
> compiler. Can the compiler stay the same with and without the cache?
It should not affect it, that is a common (universal?) design goal for a
cache (at least until RowHammer!).
But it does dramatically decrease predictability and increase
complexity, violating two of the stated goals of Project Oberon.
Bernhard's suggestion about cheap hardware is very practical, and his
frustration palpable, but the bigger question is why bother with FPGAs
at all then, if cost is such an issue? And does one neglect the value
of one's time, to learn and maintain this stuff, when it's all a moving
target? That cost is FAR larger, even when an implementation already
exists.
For some people, I'd wager that the FPGA is just in practice an
immutable platform, which happens to run Oberon, the real object of
desire. If you just want such an Oberon system there are many choices
already, and cheap ARM boards exist to port to - even cheap RISC-V
boards now. So now discarding that for a moment, for an FPGA board what
would be the purpose?
Cheers,
Paul
More information about the Oberon
mailing list