Hi everybody!!!

I've just received a mail from a friend of mine... =)
He read an interesting piece of news on "Notizie Amiga", an italian website
concerning Amiga, Aros, and MorphOS news.
I won't write anymore.... I let you the link and I'm gonna keep on thinking about....

http://us.fixstars.com/products/powerstation/

Vapor(hard)ware or not? I hope it isn't....
...Ehm.. Team..?!? How are you? =)
Seriously, could MorphOS run on this beasty machine?

Me like very much!!

THX!!!

Quote:

Theoretically it could, I guess. On one core, in 32 bit mode. But you would really like to have a SMP capable OS to harvest the performance you pay so much money for...

Yes, a G5 64bit quad core (SMP) processor would be wasted on Morphos. Frankly, it would be nice to see a new low cost machine that Morphos could be ported to.

...ooohhhhhhhh....
As me and my friends were thinking, since this machine is described like the possible machine where to run MorphOS and/or AmigaOS, a waste of resources...
All of us know how light and flexible is MorphOS (remember me something...).
So mates, you are right! And yes, Jim: a low cost/high quality machine for MorphOS would be surely appreciated.
Even without all that "hard core" inside!
.......... Well, if Team should think about it...... Let me know...!!!

> http://us.fixstars.com/products/powerstation/

First mention on MZ here 5 months ago. And there have been several more since then.

> Vapor(hard)ware or not?

Of course not. Many people have been using it for more than half a year. Just search the net.

> a G5 64bit quad core (SMP) processor

...doesn't exist. It's two dual-core processors :-)

Thanks Andreas!
I didn't follow the whole topic that time and so I didn't read about it.
I know that is not a vapor hardware
I didn't clearly explained that phrases.... whoops!
I meant that system could be (is) vapor(hard)ware for MorphOS users, as explained above and in
the topic you mentioned.

I probably should have done more than gloss over the posting, myself. Nice, kinda reminds me of Apple's last PPC based G5's though.
On SMP or dual core processors, how hard would it be to rewrite the kernal to enable multi-processing? I thought that since 680XX processors were such good multi-taskers, that Amiga derived OS's would be good at working with multiple threads.
If quark is too unweildy for the job, other kernals do exist that could handle concurrent, preemptive, priority based multitasking. Since morph is so compact, how hard would it be to re-write it for another micro-kernal as long as we retain a PPC processor..

[ Edited by Jim on 2009/3/5 20:15 ]

@Jim

I don't think quark is the problem.

Rather than trying to integrate the additional processors into the OS in such a way that the OS handles the multiprocessing for you, what about making them available as co-processors? That'd be a bit of a hack but maybe better than letting them go to waste. My guess would be that specially coded video and audio apps (or anything so processor intensive) could benefit from the 'spare' cpus.

> I probably should have done more than gloss over the posting, myself.

I already told you about the CPU of the PowerStation one month ago :-)

> kinda reminds me of Apple's last PPC based G5's though.

Yes, the similarity to this one is clearly intended:

"It's been three years. The wait is over. The Power workstation is back. Not just a simple replacement, but a well designed, perfectly packaged, readily upgradable, and far, far, more open source friendly system."
http://us.fixstars.com/products/powerstation/intro.shtml

"Not even when you-know-who offered a PowerPC workstation did the community enjoy such a well supported powerhouse."
http://us.fixstars.com/products/powerstation/open.shtml

> On SMP or dual core processors, how hard would it be to rewrite the
> kernal to enable multi-processing?

Quark probably has been supporting multi-processing from the beginning.

> I thought that since 680XX processors were such good multi-taskers,
> that Amiga derived OS's would be good at working with multiple threads.

Yes, but only time-sliced, not concurrently.

> If quark is too unweildy for the job, other kernals do exist that
> could handle concurrent [...] multitasking.

It's nothing to do with Quark, which probably already supports concurrent multitasking. It's MorphOS's ABox being only a single-threaded Quark task preventing SMP for the ABox. Thus, other "kernals" wouldn't help a single bit.
At least, that's how I understand the matter.

Time slicing looks concurrent if done fast enough. I sold 68000 based computers in the '80's that could support multiple users and they multitasked quite well. As long as the latency time to service a process (signalled by an interupt or a system module) remains low, the difference between a time-sliced single processor and a multi-processor system isn't truely noticable.

> Time slicing looks concurrent if done fast enough.

Yes, that's what it's intended to look like, but it doesn't help with multi-processing, which is the context here. X times the processor would still provide x times the performance, roughly and simplified expressed.

> As long as the latency time to service a process [...] remains low,
> the difference between a time-sliced single processor and a
> multi-processor system isn't truely noticable.

That doesn't make any sense at all. If you start x identical finite tasks on an x-way multi-processing system, they only take roughly 1/x-th of the time compared to the same x tasks on a single-processing system with time slicing.