[libre-riscv-dev] PowerISA, NLNet grants

[whygee at f-cpu.org]

Hi list !

On 2020-01-18 20:36, Immanuel, Yehowshua U wrote:
> That being said, if the $5000 annual membership fee for RISCV is an
> issue, I am willing to pay it personally.
> I believe sticking with POWER will cost much more than $5000 in labor
> than sticking with RISCV.
> 
> What are your thoughts?

My personal, detached yet biased opinion :
I think it's not just a matter of money.
The NDA and attitude matter a lot.
Luke does mentions systemic problems with the RV community
and I'm looking for more than clues, to better understand,
but we must also must get something out of the door one day.

Technically, RISC-V is alluring for many reasons and we have
to wonder WHY we are/feel concerned and what we want to achieve.
POWER will probably not allow us to do it either for reasons
you explained well.

In the end, the goals are to make a privacy/safe processor
that could be widely adopted, IIRC (otherwise it wouldn't make sense).
Maybe we can just fork RV after all, regardless of the opinions
of the others, in order to do the graphics stuff,
and if it works well enough, others will merge/adapt the work
with fewer efforts than if we had to support POWER.
Forking is frowned upon by the Foundation but can they really
prevent/forbid/avoid it ? They manage the ISA but they say it's "open".

Let's take the "Open Source" stand of the RVF at their word ;-)
https://www.youtube.com/watch?v=nPXdbm9lc3A seems "inviting"
so let's see... The trick would be to say it's "RISC-V 
derived/compatible"
but not compliant, and use the word "fork" in the name instead of RISC-V
(to prevent some of the trademark problems Luke fought).
That would make it a "Fork CPU" (a funny pun, considering
I own the F-CPU.org domain)

Yes, these divagations tell me I should go back to bed
and sleep at last ;-)

yg