Is anyone looking at the possibility of a PS3 port?
There are a number of reasons Haiku would be an ideal OS for the PS3.
Yellow Dog Linux will be available shortly and Gentoo is already planning a port, others will no doubt follow.
The problem with Linux however is these days a modern Linux desktop system has pretty hefty memory requirements. When I used to run BeOS it ran fine in 32MB for the most part and 128MB was positively luxurious.
Additionally BeOS (and by extension Haiku) is heavily threaded and SMP capable, a necessity to take advantage Cell.
While Cell has something of a reputation for complexity most of that is actually due to the complexity of multithreading - something not exactly new to BeOS developers.
There is additional complexity in handling Local Stores and vectorising code but vector code is easy* and handling the LS is really about moving data about in blocks and using double buffering.
That said you only need the SPEs for compute-intensive code, usually a tiny fraction of an application. There’s plenty of papers by people using Cell and on Cell friendly apps you can expect 5 - 10X speed ups - or better - so it’s worth it if you need performance.
There’s also the advantages of a fixed platform, no need to write drives for all sorts of chipsets, gfx cards etc.
Anyone looking at or considering a PS3 port?
*I wrote my first vector code a few weeks back, I found it remarkably easy and within 3 days I had the code running 20X faster.
i see big promise into bringing Haiku to PS3. I think it should only be on it once R1 is out to not lose focus. That said it could be interesting to look into it in advance to find more problem into the portability of haiku as an OS.
One of the reason it would make a good platform is that even if the price is high, it’s gonna eventually be 100$ machine and it make an easy complete hardware to find for cheap, in always the same configuration and worldwide.
i see big promise into bringing Haiku to PS3. I think it should only be on it once R1 is out to not lose focus. That said it could be interesting to look into it in advance to find more problem into the portability of haiku as an OS.
One of the reason it would make a good platform is that even if the price is high, it’s gonna eventually be 100$ machine and it make an easy complete hardware to find for cheap, in always the same configuration and worldwide.
By the time it’ll be a $100 it’ll be exceptionally weak compared to actual computers, power-wise. I don’t see anyone left using PS2 Linux, do you? It (the PS3) is already chronically low on RAM.
By the time it'll be a $100 it'll be exceptionally weak compared to actual computers, power-wise.
I think that’s somewhat unlikely, benchmarks from several different sources show that when properly utilised Cell can be something in the region of 10X faster than a top end desktop core. It’ll be a long time before x86 processors even catch up never mind surpass it.
Quote:
I don't see anyone left using PS2 Linux, do you?
On a 300MHz processor with no access to the DVD, it’s probably not very usable, PS3 is completely different.
Quote:
It (the PS3) is already chronically low on RAM.
This is a point I made in my original comment, however it does (or at least should) not apply to haiku as BeOS was never a huge bloated system. This is why Haiku would make an ideal PS3 OS.
By the time it'll be a $100 it'll be exceptionally weak compared to actual computers, power-wise.
I think that’s somewhat unlikely, benchmarks from several different sources show that when properly utilised Cell can be something in the region of 10X faster than a top end desktop core. It’ll be a long time before x86 processors even catch up never mind surpass it.
Quote:
I don't see anyone left using PS2 Linux, do you?
On a 300MHz processor with no access to the DVD, it’s probably not very usable, PS3 is completely different.
Quote:
It (the PS3) is already chronically low on RAM.
This is a point I made in my original comment, however it does (or at least should) not apply to haiku as BeOS was never a huge bloated system. This is why Haiku would make an ideal PS3 OS.
It might be a long time before desktop CPUs pass it out, but its also going to be a LOOOONG time before it costs less than $400 - and there’ll be more capable desktops available for that price by then.
It also has no real access to the BDROM - not gonna be any movie playback for a start.
And even on BeOS, when you throw Firefox + other apps at it, 256MB RAM starts to feel a bit constrained.
i agree the memory is no fun on PS3. That said my interest in a port is that it’s gonna eventually become a cheap “same in all places” system. That make it ideal for second computer, office computer and such. Heck, my dreamcast if it would have more RAM could still be usefull with an OS on even at a small 200Mhz.
The best would be if Sony is really serious about it, to add a memory expension to their eventual “slim” model (i expect one as the PSX and PS2 got such small enclosure). Of course that memory would only work for the other OS section of the console. Yet, at 256M of ram, this is already what i have in my PC, it’s border line for monster like “browser” but managable with HD swaping.
Even if i will not claim haiku should waste develloper on such a port, i think cell based machine will need to eventually be in the landscape anyway. So if the money to do it don’t in part come from IBM or Sony or toshiba i would let it as a possibility for now.
In the end it all come down to something i hope Koki would eventually anounce before the R1 release: A haiku officially line of hardware that get suported at least 1 year in advance and available in most country. Why? So that everybody that want to buy a computer today that is top of the line can do so knowing it’s gonna eventually be haiku compatable kinda soon or at least be a priority. That is also a way to get some buck from a manufacturer, of course not at Apple from INTEL level of discount but still.
In the end it all come down to something i hope Koki would eventually anounce before the R1 release: A haiku officially line of hardware that get suported at least 1 year in advance and available in most country. Why? So that everybody that want to buy a computer today that is top of the line can do so knowing it's gonna eventually be haiku compatable kinda soon or at least be a priority. That is also a way to get some buck from a manufacturer, of course not at Apple from INTEL level of discount but still.
IMHO. I believe this is the wrong path to go. I believe Haiku’s core value should remain the development and progression of a great OS. Getting involved in the hardware sales market is off the path to that core value and in many ways reminds me of what went wrong with Be.
sorry i might have badly expressed myself. I don’t mean selling hardware, just make an official “statement” of a target hardware.
Once a haiku user buy such a system, he could email both haiku and the laptop maker (or an independant site that do stuff like petition perhaps posting the serial number of the machine). That way, if it eventually reach 1000 (i think it’s possible number if the laptop is 100% supported), OEM might start to notice and perhaps give some money or at least better doc to haiku to be able to make the next update of supported list.
That is not possible on a store basis, too impratical, but at a constructor level the added sale + exposure might be worth some buck.
Another benefit of such a ploy would be to slowly aggregate, over time, haiku user PC upgrade cycle so that most wanted driver could be common to more user.
If Haiku becomes popular along the level of BSD, etc., then you will have projects spring up that will port it to other platforms like the PS3. This is what has happened with Linux, BSD, and the rest.
If Haiku becomes popular along the level of BSD, etc., then you will have projects spring up that will port it to other platforms like the PS3. This is what has happened with Linux, BSD, and the rest.
So, just sit tight.
Agree. This is the nature of open-source operating systems. They evolve thanks to other users/corporations that might need or want the system on a specific platform.
They might even get into the svn-tree someday.
Lets just focus on one thing at a time. Such as R1.
I’m really conflicted about this. It could take away the whole focus on the desktop and everything. But one of our (and that includes syllable and skyos) problems is the lack of hardware. We just throw our OS to everyone and they complain it doesn’t work. oh the HCL is too small, won’t work with the new card, etc.
And then we hear the slashdot chorus: oh it’s not GPL oh it’s not like linux, why bother?
But if we can get this on PS3 (more likely future versions) that would bring up the OS to a lot of people because a console is very simple. Everything will just work and there’s no Windows to deal with. Whole new category of apps will be created.
So i really think we need to work on a relationship with Sony and Nintendo too. Bring back the concept of the home computer like the original amiga
I was under the assumption the PS3 has 512MB of RAM, 256MB available until Sony releases its linux kernel patches to access the SPEs, video hardware and the entire amount of RAM :?:
I was under the assumption the PS3 has 512MB of RAM, 256MB available until Sony releases its linux kernel patches to access the SPEs, video hardware and the entire amount of RAM :?:
The hardware 3D doesn’t have a driver yet and the BluRay access is controlled (but doesn’t appear to work?) as far as I’m aware there are no other restrictions.
The SPEs can already be accessed. The video memory might also be accessible but reading it from the Cell may be horribly slow. It can be faster but only if you use the RSX to fetch the data.
Not that I’m going to do it, but just like porting Haiku to the Mac mini, it’ll happen. Someone will get the itch and it’ll happen. Ain’t open source great? :^)
