flypig (+4.6)
2/9/05 7:18PM |
This is really fantastic. In my opinion RISC OS has been crying out for an OpenGL port for far too long. I can't wait to try it out. 
If the performance is any good it could revolutionise graphics on RISC OS, and might even result in a few games appearing. |
SimonC (+1.0)
3/9/05 1:57AM |
Flyping seems to have said most of what I would say. Excellent indeed, and I can think of one or two complete vapourware plans I've got that could make use of this (but never will, knowing me). |
hEgelia
3/9/05 9:47AM |
"Meanwhile, the source to Quake 3 was recently released for free."
As if... 
I wonder if Castle got in contact with Simon Wilson about his work. |
jess (+1.0)
3/9/05 12:02PM |
This is excellent news.
Is there any indication of when any demo will be released? (Like the teapot mentioned in the previous article) |
AW
3/9/05 12:30PM |
If anybody wants Quake then JL from AcornArcade is as familiar as anybody with the enhanced versions of the game and the sequels in terms of graphics acceleration.
A recent version of Quake running on RISC OS would be a great eye-catcher for the Iyonix. |
ksattic (+9.2)
3/9/05 6:39PM |
I've just uploaded a spinning earth demo, which the adventurous can recompile and link to use the included moon texture. The moon texture is nearly 8 times larger than the earth texture, and yet the framerate is unaffected (512x256 compared to 1440x720). This is due to the impressive fillrate of the GeForce card. The download for the libraries includes a cubes demo, which demonstrates simple mouse interaction.
The teapot relies heavily on triangle rate, which isn't so good right now. The driver does not take advantage of all the pipelines available, so is bound by the lack of floating point in the Iyonix. Hopefully this will be addressed soon one way or another, perhaps by utilising the GeForce more for transformations, or writing a fixed point driver. Any help here would be appreciated.
Castle have been in touch with me and I am working with them to get 3D acceleration working on later GeForce cards. I am saving to buy a GeForce 4 or MX to test with. |
not_ginger_matt (+4.1)
3/9/05 11:50PM |
"I am saving to buy a GeForce 4 or MX to test with."
I'm probably going to get modded down for this comment, but please don't spend any of your money funding this. Castle are a commercial entity, and really need to recognise that they must either do the work themselves or actually fund the relevant development.
As noble as it is for developers to fund the advancement of the market, this is something that can't go on forever. When was the last time a company decided to save up and pay to have your car fixed? |
dgs (-1.6)
4/9/05 12:17AM |
not_ginger_matt: "please don't spend any of your money funding this... When was the last time a company decided to save up and pay to have your car fixed?"
If Castle promised your Iyonix would have hardware accelerated 3D graphics when you bought it, then maybe you could claim that not having such facilities would make it "broken".
Otherwise, I think you should re-read your comment, and withdraw this latest foolish mis-application of a car analogy.
dgs |
jonix (+7.3)
4/9/05 1:40AM |
In reply to dgs:
I'm not sure not_ginger_matt meant it in the way you seem to have interpreted it. Certainly the car analogy was not appropriate but he has a point.
I think it would be in Castle's interest to help Simon by funding him with the purchase of such a card - but that's my opinion. |
RobA (+1.0)
4/9/05 3:37AM |
While I agree with Jonix in general, I suspect there may be an assumption of a particular contractual arrangement for the work being in place. If someone is being paid to spend time developing something using someone else's facilities, then the financial reward should be structured like that of an employee, but if the developer does the work on their own behalf and on-sells the result, then another set of rewards should apply. This is not to mention intellectual property implications, as well.
I suspect ksattic has been around long enough to know the difference, and has set up the arrangements accordingly - not that I want to know any details! |
ksattic (+4.0)
4/9/05 7:23AM |
In reply to RobA:
Just to clarify: the work I did was completely individual. I sent an email to Castle to let them know I had succeeded, and some testing was done by them on later GeForce cards. They offered to host my code on iyonix.com. In addition, I was provided a test nVIDIA module to bring up later GeForce cards under RISC OS. I'd like to have a more recent card provided to me, but in reality, I can't expect anything in such a small market.
I have a strong graphics background, so given any reasonably documented PCI-based graphics card, I'm sure I can knock something up in a few months. I'd love an MPEG decoder board if anyone has one lying around.
I've already received a couple of donations, which will definitely pave/pay the way to future greater projects! |
not_ginger_matt (+2.5)
4/9/05 5:23PM |
"Otherwise, I think you should re-read your comment, and withdraw this latest foolish mis-application of a car analogy."
I've re-read my comment, and still fail to see how you missed the actual point I was making. You seem to be simply jumping on the standard Slashdot mentality of 'oooh... it's a car analogy, that must be wrong!'
The problem you seem to be forgetting (and the one I was trying to illustrate) is that much of the current development is being personally funded by various developers who don't have big pockets. I thoroughly beleive that we need the hardware manufacturers and OS maintainers to help fund these people with their projects, even if it's simply to buy people like Simon the relevant hardware. |
dgs (+1.0)
4/9/05 5:36PM |
not_ginger_matt: I don't read Slashdot.
"much of the current development is being personally funded by various developers who don't have big pockets. I thoroughly believe that we need the hardware manufacturers and OS maintainers to help fund these people with their projects"
That would indeed be nice, but I don't think that telling Simon he shouldn't buy a GeForce 4 card to test his software, if he's decided to do so, is the right way to go about it.
Sadly, some RISC OS "hardware manufacturers and OS maintainers" don't have particularly big pockets either. (Although Castle don't necessarily fit into that category).
dgs |
guestx (+1.0)
4/9/05 6:21PM |
In reply to flypig:
"In my opinion RISC OS has been crying out for an OpenGL port for far too long."
A cursory examination of the IyonixMesa Web page reveals that an OpenGL port has been around for ages, albeit without hardware acceleration support, perhaps because there wasn't a stack of suitable hardware to support back in 2001.
If there's an "at long last" moment here, it's the availability of support for the highly-proprietary graphics hardware of the Iyonix (and the fact that the Iyonix includes something which does 3D acceleration). |
flypig
4/9/05 11:33PM |
In reply to guestx:
I was aware of the previous implemention, but although I never tried it I understand that it was old, unmaintained, incomplete, and not fast enough to sensibly support any sort of animation. I don't mean to knock the work that was put into making it available, but as you point out it's probably only the Iyonix that has made it a sensible proposition.
I take your point about the hardware acceleration, as perhaps this is the greatest step forwards. Whatever the underlying reason, it's still going to be great to have a useable version of the OpenGL API. The big question is now going to be whether developers make use of it. |
em2ac (+2.5)
5/9/05 1:34PM |
A ture milestone...I read your posts on the forum (linked in previous artical)
Lots of hour's work.....glad toi hear you finally made it! |
mrchocky
7/9/05 3:40PM |
Meanwhile, I've not looked at Simon's work in any detail, but there is quite a bit of my work and potential ports that can make use of this. I don't have many specific examples to hand, but one is XPilotNG and some other games. In time, I'll do a write up on http://riscos.blog.com/ and integrate his work into GCCSDK. |
SimonC
8/9/05 12:05AM |
Now that this is here, can anyone recommend a started on OpenGL (I think I'm right in thinking that Mesa is the same API?) |
druck
8/9/05 12:14AM |
Am I the only person having problems running the earth demo? It hangs the machine with an outline of an error box showing on screen. Error logging reveals:
07 Sep 01:14:14 000 00800E07: Error from (unknown): Internal error, no stack for trap handler: Not enough memory, stack overflow, pc = FC167384: registers at 001154E8
Unforunately Simon has no contact information on his website thats I can see, and there is no sign of a readme file anywhere. |
joty (+1.0)
8/9/05 1:07AM |
In reply to druck:
concerning contact address, have a look at the bottom of the page http://www.simonrules.com/iyonixmesa/ |
ksattic (+1.5)
8/9/05 6:40AM |
I've just uploaded a new version of IyonixMesa with red/blue swapping in place to cope with the different byte ordering of RISC OS. This means that colours rendered by Mesa look as they should without any faffing required. David - I've emailed you about your problem. |
druck
8/9/05 9:02AM |
Thanks joty, I obviously didn't scroll right to the bottom.
Thanks kstattic, I'll be in touch this evening. |
druck
9/9/05 9:22AM |
It looks like the problem was running the earth demo from a 2048x1536x32bpp desktop. Its fine at screen sizes up to 1600x1200, but at 1920x1440 the texture is missing, and higher than that it doesn't run at all regardless of Wimpslot allocated.
Is there anyone else running the Iyonix at a proper high resolution on a 22" CRT that has tried it, or has everyone gone down the lower resolution LCD route? |
mrchocky
28/9/05 5:26PM |
Here's an OpenGL program that could be interesting. Although even on an Iyonix I suspect it might be pushing it: http://www.earth3d.org/ |
| |
Simon Wilson has released the first public version of his 3D graphics driver for Iyonix users. The software library, 
