[libre-riscv-dev] Fwd: Preparations The Libre-RISCV SoC
Luke Kenneth Casson Leighton
lkcl at lkcl.net
Sat Mar 30 16:16:36 GMT 2019
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Sat, Mar 30, 2019 at 3:44 PM Aleksandar Kostovic
<alexandar.kostovic at gmail.com> wrote:
> > ok so good news here, we pass the 2nd criteria stage, now it goes
> > through to an independent review.
> Thats awesome! Congratulations to all of us! :)
> we will need a one-paragraph
> > summary of the project.
> here's mine idea quick idea(idk if it should be any longer or shorter):
shorter. so, basically, i am guessing that there will be a group of
people who will read lots of these summaries (possibly all in one
day?), and if it's anything like the EU Framework Programme that my
friend phil volunteered for (unpaid), the people doing the review will
be *very* independent, as in, they will be volunteers: just ordinary
aleksander, have you ever seen this? it is really *really* worth watching:
so i'd like to suggest covering "why" as the opening sentence, in a
way that shocks people into realising how serious the situation really
is. then we can go into "how" and "what".... it's a lot to pack into
a single paragraph.
"why" i think should be along the lines of:
"It is the year 2019 and it is not possible to buy a mass-produced
laptop, tablet or smartphone where the user can replace all of its
software, without loss of functionality. Some processors are
DRM-locked at the bootloader, whilst others have proprietary 3D
Graphics, Video Processors and NSA-spying backdoor co-processors (such
as the Intel "Management Engine"). There is no way for users to audit
and trust the hardware or the proprietary software"
"Therefore, shocklngly, the only way to restore and engender trust is
to literally make a new processor - one that is developed
transparently and may be independently audited to the bedrock."
then we can go into "what", however it has to be really really short
by this point, as it's just *one* paragraph. URLs are probably not a
good idea, much as i'd like them, with perhaps one exception,
https://libre-riscv.org/3d_gpu/ which contains cross-references and
shows that we're serious about the transparency, and is like the
gateway to a bunch of links etc. etc. which they can look at.
ultimately, though, i think they want something that makes sure that
NLnet was fulfilling its Charter by selecting us. also, it's going to
be made public, probably on the EEC website or something.
> The Libre-RISCV SoC is an ongoing development project which aims to
> fulfill the spot for libre low power System on a Chip - SoC capable of
> running full operating system. Our main goal is to design a chip based on
> RISC-V ISA, that if fully open. Both in terms of development discussion and
> code. We provide the source code of the SoC and its peripherals. The SoC
> would consist of the following features:
> - Four custom 64-bit cores based on RISC-V Instruction Set
> - GPU capable of rendering up to 1280 x 720p reolution at 25 FPS, 100
> million pixels/sec, 30 million triangles/sec and 5-6 GFLOPs of computing
> - FPU
> The source code of the SoC is written in nMigen(
> https://github.com/m-labs/nmigen), a python API for building digital
> We hope to design the chip which would fit just under 2.5 Watts of TDP,
> manufactured on 28nm process.
> The driver of the SoC is intended to be Kazan(
> https://salsa.debian.org/Kazan-team/kazan), a Vulkan API driver that
> supports cross-platform rendering, written in Rust language.
More information about the libre-riscv-dev