arm-netbook

arm-netbook@lists.phcomp.co.uk

1082 discussions

libre-riscv.org 3d gpu project
by Luke Kenneth Casson Leighton 27 Sep '18

27 Sep '18

long story, there's a new project underway which is to create a 3D GPU (and associated complete processor) entirely as a libre project, using RISC-V as the basis. there's a lot to get done: the roadmap is quite straightforward: http://libre-riscv.org/3d_gpu/roadmap/ it'll be a vulkan3d *software* driver written in rust and using llvm. where and if absolutely strictly necessary intrinsic inline assembler will be placed into the rust code: for the rest it will be straight to llvm, and from there using a hardware-level parallelism API called "SimpleV" which will take care of, at the hardware level, a variable-length "vectorisation" aka variable-size SIMD (which is a really *really* misleading way to describe it but SIMD and vectorisation is what most people know). for the most part this is a *software* project not a hardware one so if anyone has software engineering skills, documentation skills or anything else, do get in touch. mailing lists: http://lists.libre-riscv.org l.

2 4

Microdesktop v1.7
by Richard Wilbur 25 Sep '18

25 Sep '18

By the way, what is the status of the microdesktop design v1.7? Have you already invested in parts (specifically the 2x10 header) and/or circuit boards?

3 13

Oscilloscope or bust?
by David Niklas 23 Sep '18

23 Sep '18

Hi, I need help here. I was designing the power supply of my laptop and I noticed that some power bucks and step up/down converters (Which both seem to do the same thing. Please tell me otherwise.), produced a more stable power supply than others. In order to find out if they are truly suitable for suppling power to sensitive electronics I've heard that I need an oscilloscope. I currently own: 1. #222634543946 "DC-DC 10/12/15/20A 150/250/300/400/1200W Step up Step down Buck Boost Converter" 2. #152710861245 "2A DC Boost Step-up Adjustable Converter Module 3v-24v to 3.3v 5v 6v 9v 12v 24v" After learning that some guy with an oscilloscope reviewed these, I'm planning to get: #122923215542 "6 Pack MP1584EN ultra Small DC-DC 3A power Step-Down Adjustable Module Buck M2H3 6 Pack MP1584EN ultra Small DC-DC 3A power Step-Down Adjustable Module Buck" I've seen many oscilloscopes online on crowd funding campaigns. I've never been certain of which to get, if any. The real professional ones are out of my budget range of about $100. *I'll pay more if I must*, but I already did not anticipate the need to actually evaluate what should be solid products. I probably should find a solid adjustable DC power supply to test these with vs. an old laptop power pack or some batteries. What I need some of these things to do is to take a dynamically variable voltage as input (Li-Ion batteries in series), and produce a constant voltage as output. Others I need to just convert to the correct voltage from the old laptop power pack. Thanks! David

3 4

Re: [Arm-netbook] Logos
by John Luke Gibson 22 Sep '18

22 Sep '18

I would highly recommend avoiding any logos, without proper historical reference. One of the problems with traditional Linux icons, are that they are very un-iconic. It's better to be textually based, in my opinion, than to use disorienting imagery. At this point, most of the clay has already set so referencing/alluding-to other gnu projects [such as blender] wouldn't be detrimental, however ascii has a very richer history of use by "hacktivists". Referencing some historically relevant (to "hacktivism") ascii iconography either overtly or subtly, kindof more or less as a hat tip to communities which support said historical events will cause the logo a greater likelihood of being regarded as iconic. I wish I had specific examples, but possibly using cloister black font would be a subtle hat tip to anonymous for some individuals use of various letters in that font as copy-cat of L from Death Note. Another thing of note, would be that we need to be careful who we tip our hats to, to be careful of who in the future people might assume us to endorse. A font is innocuous enough, that it can be adapted later due to circumstance (should a need arise to disassociate) without damaging the recognize-ability of any logo. The universal and modular style of blender, is a good point to mimic. A solid dot in the center of the "O" would probably be a subtle enough correlation to the blender logo. This is the pattern of thinking we need in developing logos and "slogans". Thinking about what sounds catchy only correlates us with random corporate culture. We don't need to be entirely original as we have a history to fall back on. However originality might also help distinguish us from our predecessors, the last thing we want is to fail to distinguish ourselves from your neighborhood corporation. I would recommend Luke to contact Wenqing Yang a.k.a. "Yummei", using the notability of the project to attract their attention. I would like to point out that despite being a cultural figure (famous artist) in the hacktivist community they previously lauched a multi-million dollar successful indiegogo which caused them much heart ache (per their blog) over legal controversy with so-called partners and their personal admonishment that they failed to do enough. >From: Alexander Ross <maillist_arm-netbook(a)aross.me> >To: Linux on small ARM machines <arm-netbook(a)lists.phcomp.co.uk> >Cc: >Date: Wed, 25 Jan 2017 16:07:15 +0000 >Subject: Re: [Arm-netbook] Logos >this all these efforts, got me to have a play around too. :) > > i was started having a go at one idea of letters inserting into each > other kinda like a module. > > I had put E and O in side the M. it then kinda looked like/spelled Meoo > like a cat. > So what about a darn cat logo? maybe a cat playing with a eoma68 card in > its paws? > > sry for the yet another internet cat image type of suggestion, never > thought id be making one heh.

9 18

Re: [Arm-netbook] Logos
by John Luke Gibson 21 Sep '18

21 Sep '18

> The idea about a dot in the "O" of EOMA made me think of fonts where the > dot is in the 0 (zero) to distinguish it from O (capital letter O) As we > all are somehow computer related I would find such a design confusing. If it was really to be in honor of the modularity and progress of blender, one would want to make the "O" not really in any particular font, but to look exactly the shape blender logo, except without the flay or the color scheme: a perfect solid-color circle within another perfect circle. It wouldn't theoretically be confusing because the mark you are thinking of is a hollow circle in an oval. Letter-o'es and zeroes can only sometimes be confusing because in some fonts they are both written as vertical ovals, not just the zero. I'm not sure on the required dimensions of a certification mark, however, the solid-shape flag-esque convention I've noticed among popular one's, often seems to detract from the continuity of the design on the objects that bear them, making them seem particularly foreign and like they don't belong. The vaporwave-esque color palate which most of them choose only add to this affect, as if they themselves are trying to distance themselves on some subconscious level from their endorsements. As a side effect to this, it makes them harder to trust. Any symbolism of any significant body, displays historical awareness.. It's what tells people whether they can trust it, before reading or attempting to judge it. Therein lies the axiom, there is no such thing as an original thought, so, if by means of arbitrary work dedicated to make more concise and thoroughly plastered an announcement of what historical or cultural event may have inspired such thoughts and such work, people will then engage in the arbitrarily difficult work of judging whether the project is sincere or not. In other words, if we don't display more than necessary effort in our symbol that can be recognized with several magnitudes less effort than it took to generate, then no highly-self-valuing person will have any sane reason to take us seriously, because if they took every slightly promising project, which failed to create a Sybil-certificate reputation in this manner, then they would never find a single valuable project in their lifetime (obviously this is an exaggeration).

3 2

So this is kind of interesting, I thought...
by Christopher Havel 18 Sep '18

18 Sep '18

https://hackaday.com/2018/09/17/a-1-linux-capable-hand-solderable-processor/ A (barely) hand-solderable Linux-able ARM SoC. US$1 each if you buy a full reel from a questionable supplier, and the package is a 176-lead LQFP (!) with half-millimeter pitch (!!) -- but still only ~US$3 for qty/1 from reputable sources, and it is /technically/ hand-workable if you've got a temperature-controlled iron with the right tip and a particularly steady hand. I'm impressed. The bad news, though, is that it's an Allwinner part with all of the usual baggage. The /worse/ news is that it's actually ~6yrs old and plods along at 1GHz. Somehow I just don't see it running Android 9.0 Pie without melting down spectacularly -- although, it might be adequate for running eg Linux Mint or something similarly lightweight... I wonder how cheaply one could manufacture a complete all-in-one PC with one of these...? Back-of-the-envelope calcs makes me think one could compete with those ~US$100-150 Atom z8300/z8350 MiniPCs on eBay and Amazon and AliExpress (none of which have screens built in, mind you), but I've been wrong before, for sure...

2 1

Shout out to EOMA68 in new article
by Pen-Yuan Hsing 12 Sep '18

12 Sep '18

Dear list, Waaaay back near the end of 2017 I was invited to write a review of the book "The Hardware Hacker" (https://nostarch.com/hardwarehacker) by Bunnie Huang, the person behind the Novena project: https://www.crowdsupply.com/sutajio-kosagi/novena My peer-reviewed review article was finally published here yesterday under CC BY 4.0: https://doi.org/10.5334/joh.11 I mention this because I managed to put in a shout out to the "the EOMA68 line of computing devices" with a link to its Crowd Supply page. First of all, please feel free to refer to this article via the DOI link above, and share it widely in whatever forum you see fit. Secondly, I'd appreciate any/all critiques of my article that you're happy to provide. I suspect the most direct way is to use the article's built-in Disqus function which you can see in the "Discussion" tab on the article's page (again, via the link above). I am by no means a hardware expert nor a seasoned veteran of the free software movement. So your constructive criticisms would be appreciated. I just hope that this will, in a small way, help promote the amazing work Luke has been doing. Thanks!

2 1

Thoughts on Why Blender should Fork
by Jean Flamelle 12 Sep '18

12 Sep '18

This post has trivially to do with arm-netbooks. I'll say, risc architectures would be optimal for an ideal virtual machine that minimizes hardware awareness in userland, which has to do with Urbit, and; has to do with what I perceive Blender could be. Blender to me, long seemed designed for discovery learning. I perceive that as, 'the [infamous] Blender way'. Why none could properly document nor tutorial'ize Blender. Always intended as a practical toy. Never to grow-out of... Until v2.8 . Violating potential for discovery learning, with a persistent toolbar where previously only the toolbox-selecting-multitool and the window-clone-or-close-multitool persisted, remaining the only tools needing persistence. Much like the Linux kernel aims for ideal hardware implementation, Blender should aim for ideal graphics processing for any given hardware. Named blender-modes shouldn't exist; only two blender-modes should exist: "Render" (for static media) and "Engine" (for dynamic media). Likewise, as the Linux kernel seeks to maximize its hardware targets, so too should Blender in to perpetuity. Even for hardware unpractical for rendering or compiling, Blender should integrate EVM render and EVM compile, or alternatively SaaSS (but only mainline SaaSS after EVM etcetera, On Principle). Blender shouldn't integrate a programming environment designed for aspiring programmers; Blender should integrate a "toy language" as a programming environment that aims to enable critical fun and intuition where even seemingly random language assortments generate interesting affects that encourage further exploration. Like every OS should have boot-to-browser as a function, so too every OS should have boot-to-blender as a function. All the while, both interfaces should encourage discovery learning. v2.8 removes "game engine" mode, as-if to say "yeah, we're narrowing our scope, so what?" while almost imagining "competing in the here-and-now is more fun than grasping at a future we won't get to witness ourselves". CC0

2 5

Open-Source Streaming Service incentivizing Free Culture
by Jean Flamelle 01 Sep '18

01 Sep '18

So there are some fairly serious talks where someone wants to do this with my help. https://riot.im/app/#/room/#Anime:matrix.org/$1535508591230126PGsxV:matrix.… Discussions start with this link, where we moved to matrix from youtube comments and I pretty heavy handedly explained why talking about this in a private chartroom won't likely get anywhere. Basically the idea is to implement Taler in an ethereum DAO and create an liberapay like that, then tie that to a website integrating whatever existing peer-to-peer networks seem practical (peer.tube and IPFS seem like candidate). People could pool funds, delegate votes, and the delegates could then leverage those funds to commission free culture anime from actual studios like Trigger (a studio already has a Patreon) and these studios are desensitized to foreign piracy. Of course doing so while giving these studios artistic license and emphasizing that altruism now will support the model, support even larger // even more obsessed fandoms (from copying, modifying, and redistributing the art), and put your foreign audience in a position to directly gouge them by threatening to sell the anime to a japanese cable company and never sell the license to a foreign actor. There are two conversations after that comment, the first is immediately after, and the second has about 140 irrelevant messages about anime in-between that and the first. (tongue fully planted in cheek, with the word irrelevant :)) tl;dr I told them why it would take about 10 years to do alone, and they still seemed to want to do it after having their heart broken.

1 0

So this just popped up on Hackaday.
by Christopher Havel 31 Aug '18

31 Aug '18

Looks relevant to me... https://hackaday.com/2018/08/29/99-pinebook-gets-kde-neon-port/ Sorry to disturb the silence.

3 5

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

arm-netbook