[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [f-cpu] Do we really need virtual memory?

Yann Guidon wrote:
ben franchuk wrote:

I think the trick for getting the F-cpu out the door is rather
not dropping features but find out reasonable subsets of the instruction set to build up the design from. Something like you bootstrap a compiler. For example can one get by with say 16 registers , a simple incriment unit and alu unit and mmu?

what you describe it IS dropping features, and modifying the f-cpu definition.
While I am reducing features, the features are a subset of the compleate
F-cpu design. However since this is still a prototype design, you don't have as large steps to test and debug, compared to the 'full feature'

the big issue with F-CPU is that we must do the right thing at the first time,
or not at all, otherwise people will stick to the wrong version etc.
In the end, the time lost in making things right is probably the same as
the time lost in guiding people back to the "right way".
No argument there.

PS. I seem to get some spam that seems to come from the f-cpu list.
Anybody else have that problem?

To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/