[Oberon] Project Oberon, RISC-V Edition

Paul Reed paulreed at paddedcell.com
Sat Dec 5 11:09:23 CET 2020

Hi Andreas,

> For those of you who have followed the RISC-V project over at UCB
> Now we only need a ... board as a suitable target.

No, you need much more than that! :)

One of the main catalysts for the FPGA Oberon project was the 
availability of a suitable board, with the right amount of memory and 
peripherals - the Spartan 3 board from Digilent. It was almost perfect 
yet still (because hardware is hard) a lot of effort to bring it to a 
working state. Availability also in the sense that ETH already had lots 
of them.

I've been keeping my eye on this eventuality for RISC-V for a long time 
and it hasn't happened yet. And a board should be chosen, of course, 
that is also hopefully not deprecated immediately after the porting 
effort is over. :(

Meanwhile, there are lots of very cheap hardware platforms based on ARM 
which could be used easily for Oberon. Cortex-M7 chips now (finally) 
have enough on-chip RAM. And there's a lot of mileage in supporting 
aarch64, as people have already pointed out.

So a new platform is always tempting, because once established one 
expects it to be around for a long time. And there's some sense of a 
first-mover advantage, if you're competitive. But real endurance comes 
from solid portability. And portability often increases complexity, 
sometimes significantly, reducing clarity, a primary objective of 
Oberon. And so it goes around...

For many, the advantage of an FPGA system is being able to change and/or 
understand all aspects in a very reasonable amount of time. A hard 
RISC-V-based chip and board would not help here. There is a always a 
steep learning curve with new hardware these days, with multi-kilopage 
datasheets - and the complexity can hide quality issues and baked-in 

I would love to see Oberon running on all these hardware platforms, so 
that users can choose the hardware appropriate to the specific task to 
be accomplished. What an amazing world that would be. But because I know 
how hard all of this is, I wouldn't ask anyone to dedicate even a 
fraction of the time I've spent. It's a big ask. :)


More information about the Oberon mailing list