[libre-riscv-dev] PowerISA, NLNet grants

Luke Kenneth Casson Leighton lkcl at lkcl.net
Sat Jan 18 23:27:05 GMT 2020


On Sat, Jan 18, 2020 at 6:12 PM Tobias Platen <hacks2019 at platen-software.de>
wrote:

>
> On 18.01.20 18:43, whygee at f-cpu.org wrote:
> > hi !
> >
> > On 2020-01-18 17:39, Immanuel, Yehowshua U wrote:
> >> I am wondering why on earth one would wish to switch to PowerISA?
> >
> > That's a good question, it seems to be somewhat political due
> > to the RVF being impermeable to Luke's arguments.
>

it's not in the slightest bit political: it's very very simple.  for well
over three years - long before this project existed - the Founders of
RISC-V have not been listening to "outsiders".  aannnnnyyyy of themmmmm.

politics involves "manipulatiion".  there's no "manipulation" here.  there
is a goal that needs to be achieved, and after more than 18 months of not
being listened to, we need to find alternatiives.

> oh and because this project does not seem to be "member" of the
> > foundation ?
>

you need to read Section 5 of the RISC-V Membership agreement.  it is
effectively an NDA.

if we became a member it would LITERALLY destroy the project, and place us
in a position of potentially going to jail for defrauding NLNet, because we
applied for a TRANSPARENCY Grant, where RVF membership DEMANDS that we
operate in secrecy.


>
> > But POWER is a "sophisticated" architecture, the little I remember
> > from it is crazy complex and RV is on the opposite side of the spectrum.
>

1) we will not be implementing POWER Vectorisation.  this reduces the
complexity by almost a hundred instructions.

2) is the following "complex"? https://github.com/antonblanchard/microwatt

>
> > Both complexity and politics will be crazy hurdles to this project :-/
> >
> >> I did read that Crowdsupply considers RISCV foundation unethical


please, i am slightly alarmed to read this statement, for its
implications.  please do not confuse "crowdsupply published an update that
i wrote" with "crowdsupply's opinion".

if you have a source where Joshua, the CEO, has, on a public posting that
*he* wrote, which *i did not write*, which said that, it would be good to
provide a link to that.

l.


More information about the libre-riscv-dev mailing list