PandaBoard
Pages: 1 2
Kuemmel (439) 384 posts |
The PandaBoard seems to be out soon -> http://pandaboard.org/ A bunch of boards seem to be shipped soon to people who propose a project -> http://www.omappedia.org/wiki/PandaBoard_Early_Adopter_Program @Jeffrey → Did you order yours or hand in a proposal…or will get yours sometime soon ? Will it be difficult to port Risc OS from BeagleBoard to PandaBoard ? |
Jeffrey Lee (213) 6048 posts |
Nope. In fact, this is the first I’ve heard of the site going live.
If you want a quick-and-dirty port (i.e. only using one core) then it probably wouldn’t take much more effort than it took to get RISC OS running on the -xM. But if you want to make meaningful use of the extra core, I wouldn’t hold your breath. It’s the kind of thing that would need a team of programmers working on it (preferably full-time), and it’s not something I’m planning on looking at until the OMAP3 port is a lot closer to completion. |
Trevor Johnson (329) 1645 posts |
Regarding the extra core, without the team of programmers Jeffrey refers to, I doubt a RISC OS port would satisfy the judges of the criterion for “Potential for the project to leverage the power of OMAP4 technology & features of PandaBoard platform”. And presumably the shared source license is not as open as would be required, either. However, the Request Program (for existing projects) could be worth a shot. |
Trevor Johnson (329) 1645 posts |
A comparable consumer product will be the BlackBerry PlayBook |
Terje Slettebø (285) 275 posts |
Great news! I had intended to skip a generation when it comes to upgrades, the Beagleboard Xm, and instead aim for the rumoured PandaBoard, since I hadn’t used my BeagleBoard for more than half a year or so before the Xm became available this summer, and had expected to have to wait 1/2 – 1 year after the Xm for the PandaBoard, but it may seem we could get it sooner… :)
Note that it says: “Selection criteria may include:” (my emphasis) As for the “coolness factor” they mention, I can think of no better project than RISC OS, but that’s me… ;)
This may be a sticking point, however… :/
Agreed. Maybe they would be lenient when it comes to the license, as well… For me, the board holds attraction even without the ability to use the extra core, namely the faster processor and higher frequency, and not at least, higher screen resolution… (finally!) I’ve got a 1920×1080 monitor, but the BeagleBoard can only run it at max 1280×1024. The PandaBoard reportedly supports full 1080p resolution. |
Jeffrey Lee (213) 6048 posts |
But if I don’t buy one, who’s going to port RISC OS to it? :) |
Trevor Johnson (329) 1645 posts |
We could club together for you, but wouldn’t be able to help with increasing the number of hours in a day!
Some students? I’m planning to take my BeagleBoard to the local uni some time. Anyone fancy a chat at the London show about sorting out a nationwide “RISC OS roadshow” organised by user groups? |
Terje Slettebø (285) 275 posts |
I, for one, would be happy to pay my share of this, as a token of gratitude for the work already done on the BeagleBoard port. However, Jeffrey has earlier expressed not wanting any contributions with strings attached, so to speak, which I fully sympathise with, and I guess no matter how it would be put, such a gift would probably be hard to interpret any other way than “Would you be so kind as to work on porting RISC OS to this one…?” :) On another note, one great thing about OMAP and BeagleBoard/PandaBoard is their approach to openness about the hardware details… In a world where many companies jealously guard their hardware designs, the BeagleBoard/PandaBoard are supplied with comprehensive System Reference Manuals and even circuit diagrams even available on the Internet. Acorn (like many other companies in the home computer era) seems to have had the attitude that it’s better to publish the technical details about their computers, and let people build on it, so these things have also been available for the BBC Micro, the Archimedes, and even the Iyonix, and the BeagleBoard/PandaBoard continue this grand tradition, which is very encouraging… :) |
Kuemmel (439) 384 posts |
@Jeffrey: I would give my share, too ;) ... I got to know if RiscOS really works on that before I buy myself ;) Besides the benefit of the screen resolution, I checked the Technical Reference Manuals of C8 and C9 and especially according to the instruction latencies the VFP-unit is almost the double speed on a single core and if people start using it, it would pay off. And also the out-of-order structure can speed up things without any change in software…so at first really no need for the multi-core-support, that’s as you said more something for small team of full team porgrammers… ...so…will you buy one and try porting or we have to raise fund ;) !? |
Jeffrey Lee (213) 6048 posts |
Money isn’t the issue. The issue is that a couple of weeks ago I realised I couldn’t remember what I was meant to be working on. And then half an hour later, just by writing things down as they came back to me, I’d managed to write a todo list containing almost 100 items. I dread to think how long the list would actually be if I went back through my ‘todo’ document and dug up all the entries that had been buried underneath all the other random notes and items. Although doing a quick port of RISC OS to the Pandaboard is likely to be quite easy, it’s also certain to add a plethora of new items onto the todo list – updating drivers to deal with new hardware features, investigating and fixing new bugs that appear, writing and maintaining extra documentation, keeping an eye on the latest hardware changes & processor errata to spot if they affect RISC OS, etc. All of which will make managing my todo list even more complex, and slow down some of the other work that I could be doing (e.g. if I hadn’t spent the past few months chasing bugs and getting the Touch Book/-xM working, I could have probably finished off the VFPSupport module, had a quick play with GCC to see how much work is needed to enable VFP/NEON support for RISC OS, etc. Or maybe sorted out the new GraphicsV API and hooked it up to KinoAMP so that you can play DVD-quality videos at a sensible clock speed instead of at needing >1GHz, etc. Or maybe, just maybe, doing one of the tasks that would move us a bit closer to being able to consider adding threading support to the OS - and the list of tasks needed for that weren’t even included in the 100-ish that I wrote out earlier.) So unless my todo list magically gets a lot shorter, don’t expect me to buy a Pandaboard any time soon. Having a long todo list – especially one like mine that has a habit of growing instead of shrinking – doesn’t do much to help my motivation, because it makes me feel like I’ve got no freedom in terms of what I choose to do next. Plus of course the depressing fact that it’s going to take X amount of months/years to work through it all. There’s also the secondary issue that I don’t want to buy a Pandaboard and then not use it to add PMT/multi-core support to RISC OS. If my work on RISC OS so far doesn’t even allow me to make full use of my current BeagleBoard, what’s the point of buying a newer model that will be able to use even less of the hardware? Why not focus on improving what we’ve got now? A turd is still a turd, no matter how big a rocket you strap it to. Also I’d much rather see someone else do the work of an initial Pandaboard port, or the initial work for PMT/multi-core support, just to ensure that if work does start then there’ll be someone else there to share the burden with me. So, I hope that explains my current feelings on the matter :) |
Trevor Johnson (329) 1645 posts |
Jeffrey – thanks for sharing all of this. So when/if the developers’ fund is available, it may not be a simple task to decide on priorities. There seem to be a number of lists going:
How on Earth will people (assuming they’re there) know where to start? Could the Roadmap do with some more info? E.g. perhaps along the lines of a table including:
Would this sort of thing actually help potential developers/ROOL decide, or is it a red herring? ROOL probably have some better ideas about coping, anyway. In the mean time, thanks so much again for continuing to chip away at your list(s). User interest seems to be snowballing – we can only hope that developer interest follows… ... and also hope that you give yourself plenty more holidays to do something different/have some fun! |
Trevor Johnson (329) 1645 posts |
It may be worth noting that the PandaBoard "has nothing to do with BeagleBoard.org" and that Gerald Coley "doesn’t have nor want one" |
Trevor Johnson (329) 1645 posts |
For anyone considering subsequent work on OMAP4 SGX, the Graphics & Connecvity Binaries are available but maybe not the source. |
Jeffrey Lee (213) 6048 posts |
Starting with something they want to work on is usually a good idea – whether that item is on a todo list or not. Unless they’re getting paid to work, in which case I’m sure whoever is paying them will have pointed them towards a suitable todo list. Of course that’s not to say that the todo lists couldn’t be consolidated a bit. I’m hoping that when the website gets relaunched, the bug tracker will be more suitable for managing tasks/todo lists.
There is a note at the top of the roadmap page saying that at some point in the future ROOL are going to go through the list and sort out more detailed descriptions, priorities, etc. |
Terje Slettebø (285) 275 posts |
I understand that developer time is the main constraint, here (and in particular, getting more developers to chip in). While Jeffrey has done, and is doing, “heroic” work on the BeagleBoard/Xm/TouchBook ports, we could really need more than one hero… Now, I’m not clear on who’s been working on the port(s), but I know there are more, whether listed officially or unofficially, but as is often the case with competent people, they’ve got plenty of things to work on… Myself, I’d liked to contribute, but I don’t feel I know nearly enough about the RISC OS internals at the moment to make a meaningful contribution, so my “contribution” nowadays is simply to learn RISC OS very well (at least, that’s the aim). After that, in line with “scratching your own itch”, as so often drives open source development, I may try to develop USB drivers for hardware I’ve got, as mentioned in another thread. For me, there’s no rush with the PandaBoard, and with only one “main” developer working on the port, I think it makes sense to continue to focus on the BeagleBoard and related hardware. If anyone wants to go for the work on porting RISC OS to PandaBoard, they are more than welcome to go ahead… :) |
Uwe Kall (215) 120 posts |
I second to this. The whole project started with Revision B6 of the BeagleBoard, and still a lot of things remain unsolved – or allow for great improvement, like VFP. So we can also wait for the panda or any equivalent board to reach that level and at least, it should have SATA. Maybe 2012? On the other hand it was good to show that new Hardware like the XM can run RiscOS, even before it can be bought in greater numbers! This opened the way for poeple who want to assemble Beagleboards into complete RiscOS computers, which might eventually broaden the user base and will then help in sales of software too. Also, I fear, that if the next generation follows too fast, everybody will be waiting for the next, better system to arise and existing users will be unhappy with their system because it works no better than, say, the A9.
Jeffrey, do me a favour and dont feel obliged to solve anything of that todo list that you do not really want. Take all the freedom you want (I am really sorry I brought up the OTG/Host port issue – I was just so frustrated to get stuck there). I do not think that anyone in the forum has any right to tell you what to do next :-) |
Martin Bazley (331) 379 posts |
Wait, what? |
Jeffrey Lee (213) 6048 posts |
[...]the website gets relaunched[...]Wait, what? Well, maybe not relaunched, but at some point it’s due to get all the backend systems updated to newer versions of the respective packages, to fix all the horrible memory leaks that can make the server fall over in the evenings if the site experiences heavy traffic during the day. Plus hopefully some fancy new stuff to make the wiki & bug tracker nicer to use :) |
Martin Bazley (331) 379 posts |
Oh. I was really hoping we might switch from Textile to BBCode there. :-) |
Terje Slettebø (285) 275 posts |
It could seem from the context that he was talking about a beaglebord.org mail address, not the PandaBoard: Måns Rullgård: Mark Lazarewicz writes: I have one of those, and I do not work for TI. Interpret that however you wish. Gerald Coley: Well, I work for TI and I don’t have nor want one. Interpret that as you like as well. |
Trevor Johnson (329) 1645 posts |
On re-reading, I agree with you. Now, what’s the chance of me spotting such a mistaken assumption in Norwegian! |
Terje Slettebø (285) 275 posts |
I just thought I’d mention that they are now apparently taking orders: http://pandaboard.org/ The price is $174, including free shipping to Europe. |
Trevor Johnson (329) 1645 posts |
And there should be a video available soon of next week’s presentation in Cambridge. |
Trevor Johnson (329) 1645 posts |
Maybe we’ll see a Touch Book board revision or v.3 |
Stephen Leary (372) 272 posts |
Jeffrey, Could we get these tasks divided up between the active devs so they get done faster? S |
Pages: 1 2