[libre-riscv-dev] xdc2020

Luke Kenneth Casson Leighton lkcl at lkcl.net
Fri Jul 3 22:37:08 BST 2020

On Friday, July 3, 2020, Jacob Lifshay <programmerjake at gmail.com> wrote:

> On Fri, Jul 3, 2020, 12:43 Luke Kenneth Casson Leighton <lkcl at lkcl.net>
> wrote:
> > On Fri, Jul 3, 2020 at 7:50 PM Jacob Lifshay <programmerjake at gmail.com>
> > wrote:
> >
> > > do we want to submit a presentation at xdc2020 which is in september?
> > it's
> > > now online and the submission deadline has been extended 2 weeks.
> >
> > good idea:
> > https://xdc2020.x.org/event/9/abstracts/
> >
> > i've added it here:
> > http://libre-soc.org/conferences/
> >
> > what topic?
> >
> how we're building an open-source GPU?

libre :)  it's not called open-soc

>  tell our story.

where to start from? there's several possible angles.  a purely technical
one could literally take 2 days to go over everything :) it took me 7 hours
straight, to explain SV to Alain!

the "why" angle, it is hardly technical at all, and can be covered quite
briefly. link to the article on the Intel Valve/Steam collaboration to
illustrate the debugging effectiveness.

another possibility: why use python (because of OO techniques) that should
definitely go in, i feel.  less code, more effective.

and the flexibility, the example of opaque object rendering requiring
*double* passes because proprietary libraries use direction-facing
detection to exclude triangles that are "behind".

what else.... there's so much but i do not want us to overwhelm people from
a software conference with hardware design details

 we should try to
> keep our issues with people in riscv out of it,

well, that is difficult inasmuch as it is part of the story as to how we
got to where we are.

*how* it is told is a different matter.  currently i have it down to 20
minutes in full detail (explanation of Trademark law if people do not
understand it) and the summary is 30 seconds.  except it is 30 seconds
which involve the phrase "anti-trust lawsuit" which makes people very
nervous if they are not expecting it (or if they are the target *of* that
lawsuit, it makes them *really* nervous)

since those are in the past
> and rehashing them will just alienate more people.

only those that do not want to accept reality, and they are already
alienated.  we know that people in RISCV are uncomfortable with reality,
and do not accept it.

if the story is told in a "complaining and whining" fashion then that *is*
guaranteed to alienate people.

the free software community already knows, now, that RISCV is Fake Open
Source.  there is no need to remind them of this reality in an "aggressive"
manner.  the package support from distros is already declining rather than

> we could also pick a different topic

any ideas on that, anyone?


crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

More information about the libre-riscv-dev mailing list