PS3 Open Platform - some big news

[Colourless]

Ew, IFF... you know who invented IFF... Electronic Arts... why would you want to use something made by EA

Seriously though, IFF, and it's byte order reversed derivitive RIFF, are decent for what they do and are still used quite a bit. They are of course the origin of good old FourCC... ok maybe not entirly, but they made them popular

 

[-tkf-]

were are all the old Amiga Demo coders these days?, perhaps they can put a few days in on this spe Gfx problem and show these new PC devs how it can be done with 256k of code....

Some of them are making games.. Anarchy Online is(was?) headed by a former Amiga Demo Coder.. awesome one as well..

The Scene is still alive and kicking.. alot!

http://www.scene.org/

 

[habbe]

Is there any projects on any type of HTPC soft for ps3 built on linux ?.

Like an little kernel booting with an nice front end to view pictures,movies and music, Wouldn't be impossible to fit on an 512mb usb stick ?.

 

[Shifty Geezer]

Is there any projects on any type of HTPC soft for ps3 built on linux ?.

Yes. It's called XMB.

 

[habbe]

Yes. It's called XMB.

Yes that would have been the best solution if it supported several movie formats and Upnp/dlna support and so on..

 

[Shifty Geezer]

There was a comment from one of the Terrasoft chaps about what application(s) to use for running PS3 as a full Linux HTPC. A quick Google didn't find something, but I'm sure someone will appear with the relevant link.

 

[legendCNCD]

Some of them are making games.. Anarchy Online is(was?) headed by a former Amiga Demo Coder.. awesome one as well..

The Scene is still alive and kicking.. alot!

http://www.scene.org/

OT: Yes we're alive

Snowblind maincoder is afaik ex-cryptoburners, IGI and Hitman.. Battlefield, black&white, max payne.. to name a few games.

 

[Titanio]

It would be interesting to see how far homebrew games can go graphically using Cell alone. For most homebrew-type-games the PPE alone would be plenty of power for 'everything except graphics', to be honest..you'd probably have more than enough room left over to accomodate the complexity of most homebrew games even if using all the SPEs for graphics. I imagine you could make some quite pretty and impressive stuff..at least it would be interesting to see, and to see some less common techniques employed. So I do hope some people explore that. But the problem is coding all of that..Panajev is right, most people don't want to be reinventing that wheel again. Perhaps there'd be more motivation for those wishing to explore rendering techniques that wouldn't fit too well on a GPU anyway. There's definitely an opportunity for a team whose interests lie in software rendering to come in and provide a library for others to use in projects more generally too...whether you have a GPU or not, most people definitely need an API they can throw their data at and forget about if they're going to write games. But who will do that groundwork, I don't know. People may be biding their time to see if Sony will unlock the GPU going forward before expending too much effort there, since such a move would render that kind of work slightly less useful..

I'm beginning to wonder if Sony is restricting GPU access in a sense to focus people's efforts on Cell. That's the main reason for them offering Linux afterall, to encourage Cell programming and the development of Cell expertise. If people had GPU access they'd possibly be less motivated to explore Cell as thoroughly for multimedia or graphics applications, and I guess Sony wants people to do that.

Games aside, I think for certain other classes of application I still want to be developing on PS3/Cell. For homebrew games, the 360 setup is definitely a better and easier entry point right now.

 

[Arwin]

Games aside, I think for certain other classes of application I still want to be developing on PS3/Cell. For homebrew games, the 360 setup is definitely a better and easier entry point right now.

True. Though also significantly more expensive. 99$ per year. But it does get you good support, it seems.

If you get that, do you also still need Live Gold, or is Silver enough?

Anyway, some people have started here:

http://forums.ps2dev.org/viewforum.php?f=27&sid=be0746f42ab2c73b1872c44039ee316e

 

[Titanio]

True. Though also significantly more expensive. 99$ per year. But it does get you good support, it seems.

If you get that, do you also still need Live Gold, or is Silver enough?

Anyway, some people have started here:

http://forums.ps2dev.org/viewforum.php?f=27&sid=be0746f42ab2c73b1872c44039ee316e

Not sure on the Live subscription requirement.

MesaGL is there and useable on PS3, but I'm not sure what kind of quality/performance you'll get out of it as it is currently (unoptimised, just running on the PPE..not even sure if it using VMX!). It might be enough for some needs, but it'd be preferable to have a library that leverages SPEs and could scale over different numbers of SPEs. I don't think the issue is hardware so much as having that library there. 'Cell graphics', depending on how well it is exposed, should probably be more than enough by most homebrew game standards, and perhaps quite surprising in the right hands, but homebrew game programmers by and large probably don't want to be coding that from scratch. It'd be nice if there was a dedicated team out there specialising in software rendering or alternatives, and if they turned their attention to PS3/Cell (even if the result was a commercial product!).

edit - for example, it would be nice if something like SwiftShader was optimised for PS3/Cell. There's another company who was working on software rendering like this also, but their name escapes me right now. I think these are very commercial/expensive solutions, though..but that asides, these would kind of be the ideal in the absence of a GPU, I guess.

 

[Crossbar]

Games aside, I think for certain other classes of application I still want to be developing on PS3/Cell. For homebrew games, the 360 setup is definitely a better and easier entry point right now.

I tried reading up on the "free" (or $99 a year for 360 develpment) XNA, but I could not really find out if you will be able to freely distribute the binaries for the 360 that you create using their tool set.

Does anyone have information about this?

 

[Titanio]

I tried reading up on the "free" (or $99 a year for 360 develpment) XNA, but I could not really find out if you will be able to freely distribute the binaries for the 360 that you create using their tool set.

Does anyone have information about this?

I don't think you can distribute binaries directly from one 360 to another, if that's what you mean. But you can share source code and rebuild etc.

I wasn't examining things from that perspective, just pure development rather than distribution and subsequent concerns. In that regard, a totally open model is obviously better, sure.

 

[Arwin]

I don't think you can distribute binaries directly from one 360 to another, if that's what you mean. But you can share source code and rebuild etc.

I wasn't examining things from that perspective, just pure development rather than distribution and subsequent concerns. In that regard, a totally open model is obviously better, sure.

Right now you can only share binaries and/or source code among other XNA Subscribers (Creators Club), at least as far as I understand it. It will be interesting to see whether there are going to be more XNA Subscribers or PS3 Linux installers, in that respect.

I assume though that if XNA takes off, then some of the better stuff will be offered through Live Arcade eventually.

 

[Titanio]

Right now you can only share binaries and/or source code among other XNA Subscribers (Creators Club), at least as far as I understand it. It will be interesting to see whether there are going to be more XNA Subscribers or PS3 Linux installers, in that respect.

I assume though that if XNA takes off, then some of the better stuff will be offered through Live Arcade eventually.

On the first point, yeah, that is correct. You don't have an open market. On the latter, though, I think they have already cherry-picked one title to release on Arcade. It's a homebrew game that was ported over from PC, I think, meant to be released sometime in the Spring.

Potentially this is all such an incredible opportunity for independent developers to see some success and make some money in a way that's been difficult to do since the early-to-mid 90s. It makes me really wish I could set aside the time and pull something together. But it's not quite perfect yet anywhere..I guess nothing ever is..

 

[archangelmorph]

On the first point, yeah, that is correct. You don't have an open market. On the latter, though, I think they have already cherry-picked one title to release on Arcade. It's a homebrew game that was ported over from PC, I think, meant to be released sometime in the Spring.

Potentially this is all such an incredible opportunity for independent developers to see some success and make some money in a way that's been difficult to do since the early-to-mid 90s. It makes me really wish I could set aside the time and pull something together. But it's not quite perfect yet anywhere..I guess nothing ever is..

Still you should do what i'm doing at the moment:

- Pull something together using the XNA framework and get it running on the PC side at least (this doesn't require a subscription..)

- Once it's done spend time refining and improving your game for as long as it takes for the following to occur..:

a) MS to sort out policies reagrding IP ownership etc..
b) MS to allow users to distribute binaries instead of source code (at the moment there's no support for binary distribution, whether u have a subscription or not..)
c) MS to establish a platform where homebrew games can be distributed to non-subscription users, either by XBLA augmentation or something similar..

Until that happens you can sit tight and smile in the knowledge that ur sitting ontop of a potential gold mine in whatever game you make (provided it's any good)..

 

[Arwin]

Potentially this is all such an incredible opportunity for independent developers to see some success and make some money in a way that's been difficult to do since the early-to-mid 90s. It makes me really wish I could set aside the time and pull something together. But it's not quite perfect yet anywhere..I guess nothing ever is..

Yeah, I agree. Also that both solutions are getting there, though. However, for me at this point in time, the Linux option to me seems the preferable one. Non-coders still have some incentive to install Linux, you can use anything that is already out there, you can play around with lots of stuff that is available for Linux in general, and it's all in - you don't need a separate PC for coding.

It is just shy of becoming perfect - as soon as they open up the RSX, we're there. Although it is a shame that it's not yet available, I think it will be soon enough. It means that I can use my PS3 to develop Linux stuff in general, but also develop stuff optimised for Cell, and eventually hopefully also using the RSX. That should get me far enough to (co-)develop something that can then be made available to all PS3 owners through Sony.

I have some experience with homebrew PSP development now, and that already works quite well - and Linux on PS3 is already a step up from there (although we did get OpenGL to work decently on the PSP with hardware accelleration ).

But I'm not giving up hope just yet - the PS2 had good support for this kind of stuff, and on that platform too, not everything was opened up on day one. Also we shouldn't forget that the PS2 was released first in Japan, so by the time it reached the other markets it had been out there already for a considerable period of time.

In this case though I'm fairly impartial - I would love MS's XNA project to succeed as well, and for it to become as good as possible. Although sometimes a bit bloated, in my daily work I exclusively work with Visual Studio .NET, and am loving it.

 

[Crossbar]

Yeah, I agree. Also that both solutions are getting there, though. However, for me at this point in time, the Linux option to me seems the preferable one. Non-coders still have some incentive to install Linux, you can use anything that is already out there, you can play around with lots of stuff that is available for Linux in general, and it's all in - you don't need a separate PC for coding.

It is just shy of becoming perfect - as soon as they open up the RSX, we're there. Although it is a shame that it's not yet available, I think it will be soon enough. It means that I can use my PS3 to develop Linux stuff in general, but also develop stuff optimised for Cell, and eventually hopefully also using the RSX. That should get me far enough to (co-)develop something that can then be made available to all PS3 owners through Sony.

I agree, I wonder if it would be possible to open up an OpenGL ES api running protected within the PS3 core OS? That way perhaps Sony could get sufficient control of the use of RSX. Prohibiting developers from unlocking disabled shaders etc. or whatever they are afraid of.

However, if you want to create shareware or freeware, Linux on PS3 does not really have competition from Microsft XNA, but I agree that XNA probably has the edge when it comes to graphics performance with the current lack of RSX support. But, personally I think C# of XNA is a little bit too high level if you want to write fast code, but I respect that some people see other qualities as more important.

I was a bit curious of the definition of homebrew, if it really included the development of programs that you cannot distribute on your own, so I looked it up at wiki and someone obviously reckoned the XNA tools as homebrew tools. The Linux environment of the PS2 was mentioned, but not the PS3. Maybe some of you distinguished gentlemen with knowledge in the matter could bring wiki up to date?

 

[Shifty Geezer]

I'm beginning to wonder if Sony is restricting GPU access in a sense to focus people's efforts on Cell.

I've always held as much. Cell as a single CPU solution for processing and display driving is a nice, cheap incentive, but needs software techniques. So through out there a Cell only platform and get people developing those techniques. You'll get an interesting mix of solutions, from which you find the one that works best in your TV/Mobile Phone/PSP2 etc.

Certainly for game development, a common framework would be ideal. It'll require the community to create though, and I question how many people are really that keen on writing a software rasterizer? More likely those that are interested want a fancy renderer, and we'll see uncommon solutions that perhaps aren't suitable as a common framework or graphics engine from which games can be developed.

 

[Arwin]

The Linux environment of the PS2 was mentioned, but not the PS3. Maybe some of you distinguished gentlemen with knowledge in the matter could bring wiki up to date?

I just made a very quick initial addition for the Playstation 3. The 360 is briefly mentioned, but should definitely get its own section also.

 

[Panajev2001a]

I've always held as much. Cell as a single CPU solution for processing and display driving is a nice, cheap incentive, but needs software techniques. So through out there a Cell only platform and get people developing those techniques. You'll get an interesting mix of solutions, from which you find the one that works best in your TV/Mobile Phone/PSP2 etc.

Certainly for game development, a common framework would be ideal. It'll require the community to create though, and I question how many people are really that keen on writing a software rasterizer? More likely those that are interested want a fancy renderer, and we'll see uncommon solutions that perhaps aren't suitable as a common framework or graphics engine from which games can be developed.

I think their current worry is the fact that RSX can read and write XDR and they are afraid people might use it to reverse engineer their security mechanism/hypervisor/etc...

Hopefully one day they will find some secure (as much as it can be) method that allows 3D acceleration through at least a high level interface that allows Shader programs to be run on the GPU and at least GDDR3 access (maybe they could allow the code to read and write from an appropriately marked XDR memory area and all accesses to even that pool of memory would be checked by the Hypervisor and terminated without warning if any kind of access is attempted in areas or ways the Hypervisor deems unappropriate).