[libre-riscv-dev] PowerISA, NLNet grants

Luke Kenneth Casson Leighton lkcl at lkcl.net
Sun Jan 19 01:53:52 GMT 2020


On Sat, Jan 18, 2020 at 9:45 PM Immanuel, Yehowshua U <yimmanuel3 at gatech.edu>
wrote:

> > So do we all agree that the trademark is not a real problem ? :-)
>
> Yes more or less… Or at least not THE problem
>
> > Some are already looking at POWER... but with a dual ISA
> > (or "why make it simple when it can be complex ?")
>
> I think we should stick with RISCV if its just a trademark issue. Like
> Staf, I also have a few connections.
> My friend bumped into the initial engineer for RISCV at the Chips
> conference and you’re always making new connections at a Conferences - so
> there may certainly be a way to angle this simply by talking to the right
> people.
>

please understand that even SiFive's own employees are terrified of
speaking out in any way that could be perceived as a "threat" to the
dominance of the RISC-V Founders (Asanovic, Patterson).

there is empirical evidence (based on timing of events) which tends to
support a hypothesis that at least one SiFive employee has been fired for
asking questions on our behalf.

It shouldn’t be hard to find parties sympathetic to our cause within the
> RISCV foundation. Often times, its more of a red tape thing. For example, I
> know Stanford has strict licensing and trademark restriction on things
> created from Academia… It may be the same case within Berkeley - but again,
> I’ll can see who I can talk to depending on IF Libre GPU is still
> interested in RISCV.
>

first, like i said to staf: please try to understand all the issues, first.

second: i do not make "final" decisions.  see Bill of Ethics (summary:
"certainty is a *pathological* state of mind).  therefore, yes, of course
it is of interest.

it requires that the RISC-V Founders *open their minds* - something that
they have not done in 4 years, *long* before i was involved.


>From the responses so far, it seems that the general consensus is that
> staying with RISCV would be easier?
>

given that all of the ISA development, and the simulator, was done based on
augmentation of RISC-V, converting to POWER is a major time / financial
setback of at least 6 months lost work.

for that reason - and that reason alone - RISC-V would be "easier".

l.


More information about the libre-riscv-dev mailing list