[Libre-soc-dev] OpenPOWER Summit NA 2020 I See you soon!

Cole Poirier colepoirier at gmail.com
Thu Sep 10 23:33:53 BST 2020

On Wed, Sep 9, 2020 at 12:13 PM Luke Kenneth Casson Leighton
<lkcl at lkcl.net> wrote:
> (do remember to trim context, cole)

Trying, still learning where to draw the line.

> On Wednesday, September 9, 2020, Cole Poirier <colepoirier at gmail.com> wrote:
> >
> > privacy or any other rights. Something we in doing our LIBRE chip are
> > actively working to fix.
> well sadly we can't do anything about the web services


> i can't quite get over the fact that
> nobody else has noticed quite how serious it is for apple to literally own
> the entire stack, from processor to product to bootloader to OS.

Yes it has quite a lot of implications, none of them good...

> > But the whole idea of "your rights are
> > whatever we say they are as you 'chose' to agree to them by using our
> > service is draconian in the extreme.
> this kind of polarisation and ostracisation was precisely why i continued
> the samba network reverse engineering i started doing back in 1995: to
> allow the windows and UNIX worlds to talk to each other without hindrance.

Very cool.

> the trick in this case (the processor) is going to be to reach a technical
> level sufficient for the priducts they go into to be "on par" with the
> alternatives by ARM, Intel, Apple and so on.

Indeed. Although the projected performance numbers for our Fall 2022
28nm quad-core 720p 3d games and 720p video chip at only 4mm^2 and $4
give me a lot of hope that we will be able to achieve this long-term.
I'm also optimistic that we can do better in certain areas. Some of
those areas are pretty obvious (security, privacy, i.e. the primary
purpose of our project) but others are TBD and I'm excited to see both
the knowns developed and the unknowns become known and developed :)

> which is an extremely ambitious goal as it means having to not just create
> a processor it means having mass volume products that *use* that processor.

Yes, this does feel like a 'moonshot' project, but thankfully you
started Libre-SOC due to the experiences and knowledge you gained from
the EOMA68 SBC and many others, and therefore even back in 2018
(2017?) you were already working on important aspects of this such as
pinmux.py, and now you continue to reiterate how important getting the
chip into mass volume products is/will be.

All exciting stuff :)

p.s. re the moonshot level of ambitiousness, I strongly agree with Yan's mail:
The goal is ambitious but necessary, indispensable even,
as you have shown.



More information about the Libre-soc-dev mailing list