| 00:00.19 | *** join/#arm-netbook mwcampbell (~matt@ip68-102-32-118.ks.ok.cox.net) |
| 00:08.40 | mwcampbell | Has anyone investigated the reasons for the speed difference between Parabola and Devuan, which lkcl reported in a recent update? I wonder how Fedora compares. |
| 00:32.35 | lkcl | mwcampbell: looking around, the archlinux-arm key developer would like more hard info from a benchmark program |
| 00:32.47 | lkcl | eliminating x11 from the equation entirely |
| 00:33.34 | mwcampbell | Maybe you should give UnixBench (https://github.com/kdlucas/byte-unixbench) a try |
| 00:34.10 | mwcampbell | The trick though is to compile it with each distro's CFLAGS |
| 00:36.06 | lkcl | mwcampbell: exactly. plus compile it with the CFLAGS being used with libreoffice... which kiiinda defeats the whole object of the exercise unless we're testing a null hypothesis |
| 00:37.16 | mwcampbell | I'd be curious to see how the A20 scores on UnixBench on any distro. Haven't bought an A20 board yet. |
| 00:37.44 | mwcampbell | I'll get a Cubieboard2 in the next few weeks |
| 00:39.42 | mwcampbell | Do you know how the I/O performance of the NAND flash on the A20 card compares to a microSD card? |
| 00:41.34 | *** join/#arm-netbook Emulatorman1 (~andre@r167-60-218-60.dialup.adsl.anteldata.net.uy) |
| 00:42.40 | lkcl | mwcampbell: haven't really tested NAND as i'm not totally happy / comfortable with it |
| 00:48.12 | mwcampbell | I find that surprising, because the usage scenarios described in the white paper and elsewhere seem to assume that the OS is stored in the CPU card's internal flash storage, not a separate microSD card. Is that just where you eventually want to be? |
| 00:50.35 | mwcampbell | Also, the FAQ on the Crowd Supply campaign page specificallya ddresses the durability of the internal NAND storage. So I figured you were already using it. |
| 00:51.32 | lkcl | mwcampbell: it's on my laarge TODO list to move to NAND but right now i'm booting from usb FEX |
| 00:51.48 | lkcl | what i *previously* was booting from was the (GPL compliant) allwinner boot0/boot1 |
| 00:52.06 | lkcl | but my foray into setting up the NAND partitions ended up corrupting the damn bootloader |
| 00:52.34 | lkcl | i cannot be arsed to sort it out, plus running from usb-FEX allows me to easily switch between mainline and sunxi u-boot.... |
| 00:52.38 | lkcl | so much to do |
| 00:53.06 | mwcampbell | OK, thanks for clarifying that. |
| 00:54.18 | mwcampbell | Yeah, I guess all the work you're doing would be spread out among multiple teams in a Microsoft or Apple. |
| 00:54.37 | lkcl | mwcampbell: f***'n'A it damn well would! |
| 01:08.13 | *** join/#arm-netbook MistahDarcy (~crash@fsf/member/MistahDarcy) |
| 01:31.34 | SPG | Luke Leighton master 9e8210c source crowdsupply/assess_campaign.py update assessment * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=9e8210c |
| 02:45.38 | *** join/#arm-netbook ganbold (~ganbold@173.244.215.173) |
| 03:57.35 | damo22 | i heard NAND has issues where just reading from it can corrupt data :S |
| 03:57.59 | damo22 | there must be a ton of patents on it too |
| 04:42.30 | *** join/#arm-netbook mm|Bot (~modelmoti@c-73-1-201-113.hsd1.fl.comcast.net) |
| 04:42.42 | *** join/#arm-netbook mm|Bot (~modelmoti@unaffiliated/modelmotion) |
| 05:23.45 | *** join/#arm-netbook normandc (~normand2@cable-192.222.216.70.electronicbox.net) |
| 05:42.51 | normandc | Hello guys |
| 05:43.20 | normandc | Anybody available to answer a question about the micro desktop housing? |
| 05:45.46 | normandc | 60+ people logged in, not a single one of them actually monitoring: seems to be the universal law of IRC. |
| 05:47.17 | normandc | In any case: the micro desktop housing has a VGA port with I guess resolution limited to 1366x768? Is it possible to ignore it and simply connect a 1080p monitor to the EOMA68 card's mini-HDMI port instead? |
| 06:35.02 | *** join/#arm-netbook Putti (~Putti@unaffiliated/putti) |
| 06:50.31 | SPG | Luke Leighton master 4093a6c source crowdsupply/assess_campaign.py update assessment * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=4093a6c |
| 06:50.44 | lkcl | normandc: sure |
| 06:51.49 | lkcl | normandc: i've been able to get the (crudely-implemented, low-cost) D2A converters to run up to 1024x768 so far... i haven't yet (manually!) found the right configs for 1366x768 yet |
| 06:51.49 | normandc | lkcl: thanks! 1007th and 1008th pledges were mine BTW :) |
| 06:51.59 | lkcl | normandc: yay! |
| 06:52.11 | lkcl | but yes you can just use the HDMI output instead |
| 06:52.31 | lkcl | normandc: you got a Micro HDMI *TYPE D* cable kicking around? |
| 06:52.57 | lkcl | it's the ones you can get from walmart, best buy, pc-world blah blah for $9 |
| 06:53.04 | normandc | lkcl: nope, figure I can purchase one somewhere |
| 06:53.07 | lkcl | normandc: same ones for smartphones and tablets |
| 06:53.23 | lkcl | normandc: ok well it's part of the cable kit if you're interested. |
| 06:53.27 | damo22 | i pledged for a whole kit housing+card+cable set |
| 06:53.45 | lkcl | normandc: the other cable in the kit basically turns the OTG port into another USB host. |
| 06:54.01 | lkcl | it's musb so is not quiiite as good as the 2 USB2 host ports |
| 06:54.04 | lkcl | damo22: yay! |
| 06:54.15 | normandc | lkcl: I thought hard about it but decided to go for the micro desktop housing instead. Liked it to be "neat" |
| 06:54.22 | lkcl | but it's enough to plug like 1 or 2 devices in. a keyboard and so |
| 06:54.28 | damo22 | i figure one day i can 3d print a laptop |
| 06:54.32 | lkcl | normandc: cool man. |
| 06:54.48 | lkcl | damo22: well if you've got the computer card that's always an option. |
| 06:55.01 | damo22 | :D |
| 06:55.04 | lkcl | it's the computer cards that we really really needed to pass the MOQ of 250. |
| 06:55.14 | damo22 | i see |
| 06:55.28 | damo22 | what is it up to now |
| 06:55.31 | lkcl | looks like we'll easily hit 400 now, possibly even be able to place an order for 400 |
| 06:55.38 | damo22 | cool |
| 06:55.51 | lkcl | python assess_campaign.py |
| 06:55.52 | lkcl | total 82765.0 |
| 06:55.52 | lkcl | after costs 70350.25 |
| 06:55.52 | lkcl | costs 42946.0 |
| 06:55.59 | lkcl | spare: 27404.25 |
| 06:55.59 | lkcl | nres: 10000.0 |
| 06:55.59 | lkcl | after nres: 17404.25 |
| 06:55.59 | lkcl | total MOQ cards 396 |
| 06:55.59 | lkcl | total MOQ sockets 280 |
| 06:56.07 | normandc | lkcl: I know the CADÂ files for the laptop were published, but I haven't found any files for the micro desktop in your Git repo? |
| 06:56.40 | lkcl | normandc: *sigh* they're around somewhere... not in a git repo... 1sec... |
| 06:57.17 | lkcl | http://hands.com/~lkcl/eoma/microdesktop/ |
| 06:57.44 | lkcl | needs updating... |
| 06:59.09 | damo22 | id imagine a decent Lithium battery would make a laptop based on EOMA68 last quite a long time |
| 07:00.02 | lkcl | damo22: 40mAh - we're gonna find out. |
| 07:00.15 | lkcl | i'm anticipating somewhere between 8 and 10 hours with careful use |
| 07:00.20 | damo22 | noice |
| 07:00.57 | normandc | lkcl: thanks, I already had a look at that folder but didn't find the files for the lasercut wood plates. I'm thinking of designing an alternate 3D printed enclosure... But it can wait :) |
| 07:01.54 | normandc | lkcl: getting "forbidden" for /~lkcl/eoma/microdesktop/microdesktop.pdf, it's the only file I don't have access to. |
| 07:02.31 | lkcl | normandc: 1sec |
| 07:02.52 | lkcl | normandc: no i don't have the DXF files, that was done 2 years ago as a prototype |
| 07:03.07 | lkcl | arranged by crowdsupply with one of his suppliers |
| 07:03.18 | lkcl | we'll need to redo them |
| 07:03.50 | normandc | lkcl: need help? I'm an experienced CADÂ operator |
| 07:04.23 | damo22 | normandc: are there any free software cad programs that are any good these days? |
| 07:04.48 | normandc | lkcl: I can produce DXF files but don't expect me to make scad files though. :P |
| 07:05.15 | lkcl | normandc: :) |
| 07:05.45 | lkcl | i use python-dxf to make DXF files these days. then fire up librecad to check them. |
| 07:05.53 | lkcl | not a fan of using GUIs for CAD work. |
| 07:06.03 | orly_owl | what do you use? |
| 07:06.06 | normandc | damo22: I'm part of the FreeCAD community. It's a 3D parametric CAD software. It's nowhere complete yet, rather quirky, but pretty powerful. |
| 07:06.16 | damo22 | nice one |
| 07:06.54 | lkcl | orly_owl: i use pyopenscad. |
| 07:07.07 | lkcl | pyopenscad with object-orientated classes |
| 07:07.22 | damo22 | normandc: heard of antimony? |
| 07:07.28 | lkcl | pyopenscad is effectively a language translator (or auto-code-generator) |
| 07:07.36 | orly_owl | sounds good |
| 07:07.48 | normandc | lkcl: I know, we exchanged emails about 3 weeks ago. But in the engineering/design/manufacturing world, *everybody* use GUI-based CAD. ;) |
| 07:10.16 | normandc | damo22: someone talked about Antimony on the FreeCAD forum, quickly checked the website. Quite frankly, I can't see myself using nodes to create complex designs. |
| 07:10.21 | lkcl | normandc: oh yeah :) |
| 07:10.35 | damo22 | ok |
| 07:10.41 | damo22 | i'll try freecad then |
| 07:10.47 | normandc | lkcl: still mean to forward our discussion to the list... |
| 07:11.14 | lkcl | normandc: would be really helpful i think, to other people to see and dissect |
| 07:11.45 | normandc | damo22: others on the FC forum don't agree with me though. :) I believe someone made a nodes plugin for Blender. |
| 07:13.43 | damo22 | id love this eoma thing to get funded |
| 07:14.00 | damo22 | i told a few people i know |
| 07:14.01 | normandc | lkcl: maybe I'm too entrenched in GUI-based parametric CAD with a history tree that is linear. Properly displayed it's pretty easy to follow. |
| 07:15.17 | lkcl | normandc: does the editor include the following: |
| 07:15.34 | lkcl | * colour syntax highlighting so that you can use both left and right brain |
| 07:15.41 | lkcl | * NP-complete programming |
| 07:15.48 | lkcl | * object-orientated classes |
| 07:15.56 | lkcl | * exceptions (try / catch) |
| 07:16.11 | lkcl | * functional programming including lambda calculus |
| 07:16.24 | lkcl | * git revision control in some form |
| 07:16.43 | lkcl | * your *OWN* preferred editor as a replacement for the built-in one |
| 07:16.55 | lkcl | ... do i have to continue the list? :) |
| 07:17.20 | normandc | colour syntax highlighting: yes. NP-complete programming: no idea what it is, not a programmer myself. Object-oriented classes: it's fully-python compliant. |
| 07:18.46 | lkcl | NP-complete means "you can write the programming language... in the programming language" |
| 07:18.53 | lkcl | like pypy was |
| 07:19.11 | lkcl | the fact that pypy exists at all tells you that python is an NP-complete programming language |
| 07:19.24 | normandc | lkcl: git revision control: being implemented I believe. Own-preferred editor: does not seem possible, but I'm not sure. As for the rest, as I said I'm no programmer and I don't use scripting in FreeCAD so I can't say. |
| 07:21.22 | normandc | lkcl: well a whole bunch of external python modules add functionality to FreeCAD. You can pretty much do anything you want... |
| 07:23.27 | lkcl | normandc: the main thing is, i believe in the unix principle: have one tool that does a specialist job, and does it well. |
| 07:24.05 | lkcl | adding built-in editors, built-in this, built-in fucking that .... i'm fucking sick to the back teeth of programmers wasting their time fucking about with that |
| 07:24.28 | lkcl | even the openscad team are doing it - "improving" the editor |
| 07:24.37 | lkcl | "improving" the look of openscad |
| 07:25.05 | lkcl | when what they _should_ be fucking well concentrating on is improving the fucking scad to STL conversion |
| 07:25.16 | lkcl | which is where the *real* value lies. |
| 07:25.42 | lkcl | for example, they could be looking at how implicitcad does its automatic rounding on unions |
| 07:25.50 | normandc | lkcl: I've worked in design for 25 years. The UNIX principle hardly applies to CAD. |
| 07:25.55 | lkcl | and working out how to add that |
| 07:26.30 | lkcl | normandc: yyeahh.... i'm not impressed. if i want a spline surface, i'll implement it myself in python. |
| 07:26.55 | lkcl | then i know what it's doing. |
| 07:27.02 | lkcl | and why it's doing it. |
| 07:27.28 | lkcl | it's a completely different approach from a completely different background. |
| 07:27.49 | normandc | lkcl: LOL. Your brain and mine are really not wired the same way. TO me the design process is highly visual. Design simply cannot be done through lines of code. |
| 07:27.57 | lkcl | i *haven't got time* to fuck about learning an entire new line of work. |
| 07:28.05 | lkcl | it frickin well can!!! |
| 07:28.17 | lkcl | i found it incredibly easy. |
| 07:28.28 | lkcl | c = cube(10, 10, 10) |
| 07:28.34 | lkcl | u = union() |
| 07:28.37 | lkcl | u.add(c) |
| 07:28.44 | lkcl | d = cube(20, 10, 10) |
| 07:28.55 | lkcl | d = translate(5, 0, 0) |
| 07:28.58 | lkcl | u.add(d) |
| 07:28.59 | lkcl | save |
| 07:29.10 | lkcl | run |
| 07:29.26 | lkcl | the run automatically tells openscad to reload the file and display it |
| 07:29.31 | lkcl | couldn't be simpler. |
| 07:29.38 | normandc | lkcl: I understand: what I'm telling you is that the whole MCAD field isn't compatible with your way of doing things. Get back to me when you start inquiries about injection moulding the laptop casing parts. ;) |
| 07:29.53 | lkcl | i even set up a vim macro which does that. |
| 07:30.22 | lkcl | if it's not compatible, all i need to do is write a replacement for pyopenscad which.... outputs MCAD. |
| 07:30.50 | lkcl | i'm happy to learn how to do that. the primitives i'm using are all cubes and cylinders anyway |
| 07:31.15 | lkcl | if i had written it *in* openscad then you're absolutely right, i'd be fucked |
| 07:31.40 | lkcl | ... but in OO python behind a clean API? not hugely looking forward to it but it'd be doable. |
| 07:32.51 | normandc | lkcl: I really don't see how you can write to CAD exchange formats like STEP (ISO10303) through python or any other programming language. You need a B-Rep kernel like Open Cascade. |
| 07:33.52 | normandc | lkcl: while we're talking :) you wrote the following comment on https://news.ycombinator.com/item?id=12231522 |
| 07:33.56 | lkcl | eyy there we go: http://www.pythonocc.org/ |
| 07:34.27 | normandc | lkcl: rrrrrrrright IÂ had forgotten about Thomas' project. |
| 07:34.39 | lkcl | http://www.pythonocc.org/quick-examples/hello-dumb-box/ |
| 07:34.56 | lkcl | looks really trivial to deal with |
| 07:35.10 | lkcl | could probably do the translator in... mmm.... about a week |
| 07:35.13 | normandc | lkcl: now that would do it, not that OpenSCAD toy. :P |
| 07:35.40 | lkcl | :) |
| 07:35.44 | normandc | lkcl you wrote: "that's right. so what we will be doing when this gets to MOQ 2500 and above is: going to injection-molding... BUT - and this is really REALLY important - the parts will be EXACTLY THE SAME so that people can, on buying a mass-produced variant of this laptop housing, go down to a local library or go to UPS's web site and order a replacement." |
| 07:36.01 | lkcl | normandc: yep. |
| 07:40.25 | normandc | lkcl: but it isn't true. You will not be able to use the parts as is for injection moulding. There's a whole series of rules that are not the same as for 3D printing. For example, you need some surfaces to have drafts so the part can be released from the mold. Sharp corners are a big no-no. Etc. See https://www.stratasysdirect.com/resources/injection-molding/ |
| 07:41.21 | lkcl | normandc: i'm aware of these things. i'll be adjusting accordingly. |
| 07:41.34 | normandc | lkcl: ok :) |
| 07:42.39 | lkcl | normandc: why you think i wrote the python-based "spline curved box" function? :) |
| 07:45.14 | normandc | lkcl: well I was not aware, trying to read your multi-thousand line python script gave me a headache :P (remember, not a programmer here ;)Â ) |
| 07:50.01 | normandc | lkcl: going back to pythonocc... I think programming for it will be more involved than for openscad, but totally worth it IMHO. It's much more powerful... And I wonder if it can export to STL (I don't see why not). It could certainly do a much better job than openscad. |
| 08:08.21 | *** part/#arm-netbook normandc (~normand2@cable-192.222.216.70.electronicbox.net) |
| 08:09.50 | damo22 | is glad lkcl is a vim user |
| 09:59.22 | *** join/#arm-netbook putti_ (~Putti@unaffiliated/putti) |
| 10:16.41 | *** join/#arm-netbook stamina (~stamina@232-056-128-083.dynamic.caiway.nl) |
| 12:19.43 | *** join/#arm-netbook Putti (~Putti@unaffiliated/putti) |
| 12:38.01 | *** join/#arm-netbook OmegaPhil (~kvirc@88.97.77.255) |
| 12:38.01 | *** join/#arm-netbook OmegaPhil (~kvirc@fsf/member/OmegaPhil) |
| 12:45.05 | *** join/#arm-netbook Emulatorman (~andre@unaffiliated/emulatorman) |
| 15:04.20 | SPG | Luke Leighton master 4810547 source crowdsupply/assess_campaign.py update assessent * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=4810547 |
| 15:54.10 | *** join/#arm-netbook stamina (~stamina@232-056-128-083.dynamic.caiway.nl) |
| 15:57.11 | *** join/#arm-netbook mwcampbell (~matt@ip68-102-32-118.ks.ok.cox.net) |
| 16:03.48 | mwcampbell | I just noticed that the A20 card has a microSD slot in the card itself. So, why include internal NAND flash at all? |
| 16:06.01 | mwcampbell | I know that last night I was asking about the ability to store the OS in NAND flash. My primary concern was the ability to move the computer and its OS together from one housing to another. But having a microSD slot in the computer card also achieves that perfectly. |
| 16:07.43 | mwcampbell | Can the A20 boot without having any internal NAND flash at all? |
| 18:42.30 | lkcl | mwcampbell: for people who'd like to install an OS and use both SD cards for data |
| 18:42.32 | lkcl | yes |
| 18:57.59 | *** join/#arm-netbook stamina (~stamina@232-056-128-083.dynamic.caiway.nl) |
| 19:23.04 | mwcampbell | If there's a working OS install on the internal NAND, but the user inserts a bootable microSD card into the slot in the A20 card, then will the A20 automatically boot from microSD? |
| 19:23.16 | lkcl | mwcampbell: correct. |
| 19:23.40 | lkcl | mwcampbell: look at the bootrom page, google "sunxi wiki BROM" |
| 19:24.55 | mwcampbell | And can the internal NAND be reflashed by the A20 itself, using an OS running from the microSD, without having to connect to a host computer over USB? |
| 19:25.13 | lkcl | mwcampbell: yes. google it, please. |
| 19:25.22 | mwcampbell | OK I will |
| 19:26.04 | lkcl | mwcampbell: remember that the A20 is extremely common, the number of people putting out different instruction online is enormous |
| 19:27.02 | lkcl | a lot of them are clueless, for example there's forums with postings in 2016 where they have *no idea* that the USB-FEX boot process has been reverse-engineered so they're still telling people to use the windows LIVESUIT.EXE program (!) |
| 19:27.10 | mwcampbell | got it, sorry for asking straightforward questions |
| 19:27.46 | mwcampbell | I mean questions with answers already available online |
| 19:27.50 | lkcl | nono it's fine, i'm just in the middle of trying to sort out a particularly difficult boot issue |
| 19:28.07 | lkcl | tell me if you find anything and i'll let you know if it's good ok? |
| 19:28.33 | mwcampbell | I'll look later when I've actually got a board. |
| 19:29.03 | mwcampbell | But for now, it's good to know that it doesn't matter if theere's an OS on the internal NAND; it's easy to pop in a new microSD card and tinker with something else. |
| 19:29.08 | lkcl | alrightman. there'll be lots of people then so lots of people to help each other out |
| 19:29.33 | lkcl | yep. the boot rom's been uploaded and reverse-engineered. |
| 22:14.36 | SPG | Luke Leighton master 5b28276 source allwinner/a20/DDR3_review.mdwn add DDR3 review for A20 * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=5b28276 |
| 23:56.59 | *** join/#arm-netbook tuliom (~tuliom@187.74.68.229) |