Announcement

Collapse
No announcement yet.

That Open, Upgradeable ARM Dev Board Is Trying To Make A Comeback

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Originally posted by SystemCrasher View Post
    Sure, 3D is neat, but it is really optional goodie for systems administration, software and hardware development, GPS navigation, browning and so on - the most kinds of activity related to computers use I really care of.
    contemporary desktops are 3d. googlemaps is 3d. sure you can live in windows95 world, but nobody else is.
    Originally posted by SystemCrasher View Post
    If you haven't got it, when I've take a look, boot seq of Qualcomm looked like this:
    you can't get that i don't care. it works for my phone, it will work for my laptop.
    Originally posted by SystemCrasher View Post
    - Proprietary firmware stays forever and runs on primary CPU, side by side with Linux.
    you don't understand what firmware is
    Originally posted by SystemCrasher View Post
    However, boot-stage treachery and cell modem design like qualcomm is IMHO even worse than proprietary drivers
    no
    Originally posted by SystemCrasher View Post
    But other than that Allwinner A20 is really good in terms of preictable, crap-free hardware
    yeah, other than being unusable for laptop, a20 is fine.

    Comment


    • Originally posted by SystemCrasher View Post
      If pal666 thinks he could do RYF cert on Qualcomm, I'm eager to see this show.
      i care about open drivers, i don't care about some arbitrary certification

      Comment


      • Originally posted by pal666 View Post
        yeah, other than being unusable for laptop, a20 is fine.
        once again, so that is is clear: this is not a single-board design where the a20 is the sole exclusive processor under consideration. the a20 is the first in the series. there are other computer cards in the series. there will be more in the future in the series. you will be able to upgrade. the cost of an upgrade will not require you to throw away the laptop. or the microdesktop. or any other compatible devices.

        if there are any issues with that approach, please tell me how you would do it better, including a description of how long you would be willing to commit to such a project, how much of your personal money you would be willing to put into it, and how much shit you would be willing to put up with from trolls on a dozen different forums.


        Comment


        • Originally posted by makomk View Post
          Which is why people insisted not having the source for it was a GPL violation but the much less free x86 equivalent is fine.
          gpl violation is not about being non-free, it is about being derived work from gpl-licensed code

          Comment


          • Originally posted by SystemCrasher View Post
            Speaking for myself I could live with 2D on most computers, it does not requires Mali usage at all. Just display controller, which got nothing to do with Mali. Sure, 3D is neat, but it is really optional goodie for systems administration, software and hardware development, GPS navigation, browning and so on - the most kinds of activity related to computers use I really care of.
            As I've explained in a reply to another comment, it is a common trap to equate hardware accelerated rendering with 3D.

            Cheers,
            _

            Comment


            • Originally posted by makomk View Post
              All versions of the Raspberry Pi require a proprietary binary blob running on a proprietary processor on order to even boot the damn thing.
              That's the GPU, running its own firmware like most GPUs do nowadays.
              The GPU initializes the SoC.
              FYI: all CPUs around are also "proprietary processors", unless they are open hardware.

              In other (better designed) systems, low-level initialization code is in a ROM in the SoC itself, I don't see what the difference is.

              At one point the Pi Foundation were considering working around the FSF's requirements by sticking that binary blob
              FSF requirements are retarded, they are OK with a firmware if it is in a chip, but not ok with a firmware if it is loaded on boot from storage. It is exactly the same fucking thing.

              (which, it should be noted, is running with full access to RAM whenever you're using the system)
              That's due to the hardware design tho, the original Raspberry is basically a GPU SoC with an ARM core added as a passenger to deal with minor stuff.
              And the same allowed to upgrade easily the CPU without changing the SoC. As the CPU component is a passenger in that design. "app processor" they call it.

              Comment


              • Originally posted by starshipeleven View Post

                > given that people keep complaining about the lack of free MALI drivers and then DON'T PAY HIM ANY MONEY, we're left wondering "what gives?".

                I've yet to see any crowdfunding campaign for that.
                Sorry for libv, but whining on IRC or forums isn't exactly the most effective strategy.

                > it's an odd situation, the software libre world.

                No it isn't, it's just different from usual commercial licensing so most people don't know how to handle this (and I talk of developers).

                The basic fact is that someone still has to pay the developers, and most libre devs are of the more altruistic kind that will gladly work for free for an ideal.
                While this may sound good, it is ultimately stupid and leads to failure.

                Libre devs must secure money first, THEN do the work.

                I'm citing Krita as an example, that is at their third yearly highly successful kickstarter (38 thousand bucks or so for the last one) but there are others that launch crowdfundings and get funded.
                You are making a very good point. I also think that the sponsorship for developing the open source Mali400 driver needs to be secured somehow. Otherwise we are not moving anywhere.

                But one difficulty is that the crowdfunding campaign must have very clear and well defined goals. People tend to have some really strange expectations and think that the Mali400 drivers are supposed to magically fix all the performance problems in their system. And if the crowdfunding delivers a perfectly functional open source OpenGL ES driver, such people are likely to be disappointed because it is not going to make their dream come true.

                As many people already know (and lkcl just confirmed), Mali400 binary drivers have been usable on Allwinner development boards at least since 2013. There are also other popular boards with OpenGL ES drivers (such as the Raspberry Pi). Nothing was preventing anyone from developing cool and useful 3D applications for GNU/Linux. So where are all these nice applications?

                For example, if we launch crowdfunding for developing the open source Mali400 drivers (based on the excellent Luc Verhaegen's reverse engineering work) and set a clear goal for developing an open source replacement for the existing binary Mali400 drivers, would many people be interested? Let's say $200K and 1 year (a very arbitrary estimate, which surely needs to be adjusted) to get more or less the same functionality as everyone already has with the binary drivers. I'm reminding that it is possible to try the binary Mali400 drivers even now, and see what can be realistically expected. Alternatively, people can also check the Raspberry Pi board and the status of its open source GL drivers.
                Last edited by ssvb; 07-06-2016, 02:22 PM. Reason: replace "kickstarter" -> "crowdfunding" and "mali" -> "mali400"

                Comment


                • vBulletin I hate you. (unapproved post)

                  Originally posted by ssvb View Post
                  But one difficulty is that the crowdfunding campaign must have very clear and well defined goals. People tend to have some really strange expectations and think that the Mali400 drivers are supposed to magically fix all the performance problems in their system. And if the crowdfunding delivers a perfectly functional open source OpenGL ES driver, such people are likely to be disappointed because it is not going to make their dream come true.
                  Once these kinds of people were parted from their money they become irrelevant noise again.
                  As long as the project itself is successful and you hand over gadgets and stuff you don't need to care about the limited amount of people that didn't understand what was going on.

                  The issue is if you blatantly lied and scammed people, then these people are the majority, and their voices will be heard.

                  As many people already know (and lkcl just confirmed), Mali400 binary drivers have been usable on Allwinner development boards at least since 2013.
                  And for the nth time: we and everyone else know that proprietary drivers can be used, there is no need to tell us again. Closed drivers are crappy, they work but are crappy and do require hacking around their bugs. I know, because on Android it's like that.

                  And as I stated somewhere else, if I wanted to use closed drivers I'd be using Windows or Android, case closed.

                  Nothing was preventing anyone from developing cool and useful 3D applications for GNU/Linux. So where are all these nice applications?
                  I swear to God, the next time I hear someone state 3D is only for 3D applications again I'm going to eat a kitten. While it's still alive.

                  Apart from mobile games that use it for 3D, it can be used for accelerating other software like say libreoffice that uses openGL (non-ES) to do some calculations in spreadsheets and render faster graphs and other things like that.

                  OpenGL ES is a shitty embedded API for shitty embedded GPUs that aren't found in PCs so there isn't a terribly large amount of linux software using it simply because they are NOT ANYWHERE NEAR common enough for everyone to target directly for PC software.

                  Still, I know that KDE has openGL ES support. I think some other DEs can use it too for accelerated rendering. Probably some web browsers can use it or be re-compiled to use it. Libreoffice is using OpenGL ES in its Android version, so it could theoretically be adapted to use that on the desktop too if someone cares.

                  But I'm going to turn the tables on you with a shocking revelation: my main beef is running the open driver on Android (cyanogenmod and custom roms) where OpenGL ES is used and the OS is actually adapted to embedded usage anyway.

                  There it would be useful while most of the rest of linux ecosystem catches up and makes a half-decent touch interface, app-like systems to deliver applications from repositories, sandboxing and other things like that.

                  Let's say $200K and 1 year (a very arbitrary estimate, which surely needs to be adjusted) to get more or less the same functionality as everyone already has with the binary drivers.
                  Ah look, another fail because someone cannot grasp basic marketing.

                  "i want to make a windows clone for free, I ask 10000 billion dollars" is a marketing fail.
                  The general rule of thumb: "look at what does Canonical, and for the love of dog, DON'T do that" can be invoked here too.
                  Let's look at their crowdfunding for that Ubuntu phone that was supposed to be both a mobile and a PC depending on the situation (*cough*MS copied this and made Continuum for their own phones*cough*) https://en.wikipedia.org/wiki/Ubuntu_Edge

                  "It had the highest target of any crowdfunded project to date, $32,000,000 over a one-month campaign" --> no. Just no.

                  You need to act like a leech, not like Ocean's Eleven. Syphon money little by little over a long time, not rob the bank with a sweeping action.


                  So, again looking at Krita's crowdfundings:
                  15-20 k$ goal, simple objective, the rest as stretch goals.

                  Next year, repeat.

                  So that even if the people to pay 200k$ don't materialize you still have enough money to start and go somewhere.

                  Next crowdfundings you can show your work and if it isn't crap people will be more interested in paying again for its continued development.

                  Just like for Krita.


                  I'm reminding that it is possible to try the binary Mali400 drivers even now, and see what can be realistically expected. Alternatively, people can also check the Raspberry Pi board and the status of its open source GL drivers.
                  The open driver will support more than just Mali400 I hope, ARM isn't making a WHOLLY NEW design each Mali iteration, an open driver is an investment for the future.

                  As for what it can do, I have seen it on Android. It's nice and can do all sorts of nice things without murdering the processor.
                  Last edited by starshipeleven; 07-06-2016, 04:03 PM.

                  Comment


                  • Originally posted by ssvb View Post
                    For example, if we launch crowdfunding for developing the open source Mali400 drivers (based on the excellent Luc Verhaegen's reverse engineering work) and set a clear goal for developing an open source replacement for the existing binary Mali400 drivers, would many people be interested? Let's say $200K and 1 year (a very arbitrary estimate, which surely needs to be adjusted) to get more or less the same functionality as everyone already has with the binary drivers. I'm reminding that it is possible to try the binary Mali400 drivers even now, and see what can be realistically expected.
                    we do know even from the small amount of work that luc verhaegen was able to do when he was sponsored to work on this that there is room for a lot of improvement in the proprietary MALI drivers. they put a huge load on the main CPU. luc was able to put in... i don't know the full details... some form of memory-management that gave a significant and noticeable percentage performance improvement over the proprietary driver for the *small* use-case that he was actually able to test with what he'd been able to reverse-engineer up until that point.


                    You need to act like a leech, not like Ocean's Eleven. Syphon money little by little over a long time, not rob the bank with a sweeping action.
                    exactly. this is how i've managed to get this far (at all) because thinkpenguin - a small-to-medium-sized business - are able to sponsor the approach that i've taken by utilising the profits from an existing stable, respected and profitable business. now, it would be nice if there were other businesses likewise sponsoring what i'm doing, as it would accelerate what we can do and open up some more avenues. but if not, that's cool too, because the strategy i have in place is to work with what's available.

                    Comment


                    • Originally posted by lkcl View Post
                      we do know even from the small amount of work that luc verhaegen was able to do when he was sponsored to work on this that there is room for a lot of improvement in the proprietary MALI drivers.
                      How unexpected.
                      I have more experience with routers, and there it's usually the case. Heck, I even got wifi cards that run better in linux (opensource driver gives better range) than on windows (where the drivers are closed).

                      Ah, btw, did you look at router-grade SoCs? They usually lack a GPU (no duh) but the more decent ones have PCIe lanes you can attach a desktop GPU to (yeah, there are no miniPcie GPUs sadly) or other things, and should have simple boot ROMs, most do support u-boot (and are already supported one way or another).
                      I don't really know is their situation as far as not using closed blobs firmwares goes. Apart from wifi functionality (which usually does need firmware) they should work even blobless afaik.

                      exactly.
                      I still think the 150k$ you're asking are a tad high tho. I know mass-producing small runs is horribly expensive, but I would have gone with a lower goal so that even if it does not cover all at 100% you still get something at all, and moving something to stretch goals to compensate.

                      Comment

                      Working...
                      X