[Oberon] Moving oberon to RISCV?

eas lab lab.eas at gmail.com
Fri Jul 27 22:05:07 CEST 2018

Jorg wrote:-
In his career he ported his compiler to quite some processors. But instead
of porting the compiler over and over again to the myriads of processors
out there, the idea was: keep the compiler constant and make it generate
code for „his“ RISC-5 processor.
So then he used pseudo-code with a VM.
Are there any of them documented?
I did that in the 70's/80's for 8 bit microProcs,
after I had adapted the expensive HP <BASIC lang like>
desktop computer, which also did IEE488 instrument driving,
for automating test equipment.
Based on a BYTE [publication] article of PASCAL a compiler,
written in BASIC, I replaced the miriad of GoTo's by a <diagram>
representing a Finite-State-Machine.
Who remembers the <railway diagrams> representation of
The diagrams where stretched-out to show 1-node per line.
Then train <jumped> to the <action> as it passed the corresponding
node/station. So eg. [AFAICR]:
 IntrFlag <-- NAME := IntrFlag
BYTE  <-- enter/push: NAME, Offset, Type toSymTable


On Fri, Jul 27, 2018 at 3:12 PM, Skulski, Wojciech <
skulski at pas.rochester.edu> wrote:

> Paul:
> > I think the first thing would be to select an inexpensive hardware
> reference platform with some longevity.
> Not sure if it is exists. You only learn about the longevity when the
> board is gone. Consider the Digilent website. It looks like a graveyard of
> great old boards.
> You can also seek the longevity on Ebay, but the supply is limited. One
> Spartan-3 Evaluation Kit is available today on Ebay for $150. Grab it
> before it is gone! This is the whole world wide supply at this point.
> > Since this board is now sadly no longer made, the decision seems a bit
> unwise, but hindsight is of course 20-20.
> This decision was not unwise. Any such board will follow the same route.
> It is the nature of these boards. The FPGAs do not enjoy the broad market
> like ARM processors, and therefore the turn around time of the eval boards
> is unlikely to last longer than a couple years. The companies like Digilent
> will switch to the new FPGA generations and abandon the older ones.
> Which is why I designed my own RiskFive. At least I know that I have the
> longevity in my own hands till they stop making Artix chips, which will
> probably happen 10 years from now.  This projected longevity is reflected
> in the cost of the board, because RiskFive will stay low volume. Unless, of
> course, you guys decide to develop some cool apps using it, in which case
> it will become both more popular and low price.
> Thank you,
> Wojtek
> --
> Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
> https://lists.inf.ethz.ch/mailman/listinfo/oberon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.inf.ethz.ch/pipermail/oberon/attachments/20180727/367951d0/attachment.html>

More information about the Oberon mailing list