Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: amyren on November 01, 2009, 10:20:16 AM
-
I just read an announcement about Tilera's new 100 core CPU, the Tile-Gx.
Wouldnt porting OS4 to this CPU be a worthy candidate to the speculations around Hyperions secret "most amitious project"?
-
the user oriented applications are mostly sequential programs in nature, having a 100 cores won't help much for the desktop market.
quote frmo the wikipedia entry : :rtfm:
Their primary market for this new chip, expected to begin shipping in 2011, is for commercial applications such as web indexing and search engine use.
I'm sure frieden brothers are keeping their hands busy for something more useful. :afro:
-
Um.....no.
Nevermind the fact that OS4.X has no SMP support (yet.....), we Amigans would never get anything quite so....... "cool".
I see us getting a port to that CPU, oh, in about 20-30 years.
Assuming that AmigaOS is still around that is.......
[EDIT] - Damn, CountZero beat me to it and a lot more tactfully than I did, I might add. ;)
-
Brace yourself for disappointment re the "most ambitious project"
-
Brace yourself for disappointment re the "most ambitious project"
@ Tension
Amiga Anywhere for OS4.X? :lol:
-
Nobody exept Hyperion themself knows which direction they are going. But I noticed that multi-core was mentioned on the roadmap for OS4.2.
I just though that it would be cool to have a cpu for the Amiga that was ahead of the intels again.
In the published article I read, it was also mentioned that Microsoft is not going to use this CPU in Windows 7, because they are already working on a CPU with 80 cores - to be released in about 5 years.
So in a way it will be like beating M$/intel at their own game:)
If the statement of having 100 cores will not help the desktop market is true, why the 80 core cpu for windows in 5 years?
No matter if OS4 will be able benefit from all the 100 cores, it should still be much faster than any Amiga HW ever seen before.
-
There are so many threads about which HW platform might be a candidate for this "ambitious project" that we may have already discussed the answer.
I haven't read them all, but as Tension says "brace yourselves for disappointment".
Taking into account that x86 has been ruled out, I decided to Google for PPC on x86. A number of PPC processor boards were returned. These seem to slot onto x86 PCI/PCIe motherboards giving PPC compatibility.
Does anyone here have an informed opinion as to whether this would be a viable option for getting OS onto "mainstream HW"?
-
The Frieden Brothers are fairly conservative in their approach, so I wouldn't expect a major shift in direction or emphasis. I concur, everyone who isn't already in the Frieden camp will find the "most ambitious project" a great disappointment, whilst those that are in the Frieden camp will become very excited about it.
-
Taking into account that x86 has been ruled out,
When did this happen?? If true, that is really sad!! The ONLY way the Amiga could still be at all viable is if it went x86. Any other platform is a dead end.
What a waste...
-
When did this happen?? If true, that is really sad!! The ONLY way the Amiga could still be at all viable is if it went x86. Any other platform is a dead end.
What a waste...
You could argue that x86 != AMD64 ;)
-
Is there an anouncement from Hyperion mentioning the x86 has been ruled out? Didn`t seem to mention much about it at AmiWest.
-
Yes, if you dig into the posts on Amigaworld.net you can find a statement from Rigo (some kind of developer of AmigaOS, not that Simon, sorry :) ) that x86 AND Mac Mini is ruled out (this statement was made last week).
Personally if the next AmigaOS system is another 1000 dollar sub 1 GHz system you can have your next gen AmigaOS system shoved up where the sun never shines... mildly put :)
-
Yes, if you dig into the posts on Amigaworld.net you can find a statement from Rigo (some kind of developer of AmigaOS, Simon Goodwin from Amiga Format if that rings a bell) that x86 AND Mac Mini is ruled out (this statement was made last week).
Personally if the next AmigaOS system is another 1000 dollar sub 1 GHz system you can have your next gen AmigaOS system shoved up where the sun never shines... mildly put :)
My thoughts entirely.
Enough is enough.
-
http://www.power.org/news/pr/view?item_key=4dbbb4f154a284bae0f4366e8172e2a3882ad2f3
..... could be.
-
Yes, if you dig into the posts on Amigaworld.net you can find a statement from Rigo (some kind of developer of AmigaOS, Simon Goodwin from Amiga Format if that rings a bell) that x86 AND Mac Mini is ruled out (this statement was made last week).
Rigo is Simon Archer.
Just did a search on Simon N Goodwin and it turns out that he's been working for Code Masters since 2003. I wasn't aware that he'd worked for Amiga Inc between 2000-2002.
I wonder if he'd do a port of Gold Mine for OS4. :)
-
".....If the statement of having 100 cores will not help the desktop market is true, why the 80 core cpu for windows in 5 years?....."
-----------------------------------------------------------------------
Because Windows 9 will need 60 of those cores just to run :lol:
-
Using large numbers of cores represents a big paradigm shift in programming. Having played with CUDA in the last year gives a taste of how radically different you have to approach a given problem to exploit many-core processors.
-
Because Windows 9 will need 60 of those cores just to run :lol:
That's so fabricated. I have it on good authority that after the next upgrade of Windoze 7, Microsoft is changing the OS name to "Portal" (as the word Windows is "so last century") and then suing Valve for the trademarks and rights to the name Portal. Microsoft will claim they had it first.
-
Using large numbers of cores represents a big paradigm shift in programming. Having played with CUDA in the last year gives a taste of how radically different you have to approach a given problem to exploit many-core processors.
So it'll be a near rewrite for anything AOS based API?
-
oh another speculation thread, or is it a joke.this have nothing to do with a PPC, a single core is really slow with in order, if you have multithreadeds programs it run fast.See here.
http://arstechnica.com/hardware/news/2007/08/MIT-startup-raises-multicore-bar-with-new-64-core-CPU.ars
"""
The short-pipeline, in-order, three-issue cores implement a MIPS-derived VLIW ISA with a few important and peculiar features.
"""
thats great news for compiling, because compile programs as OWB, chrome, firefox is so extreme slow and need hours that really demotivate to do it.
but on desktop, most users notice that most time more than 2 cores not need.
I hope we see end of year what is ambitious project.If not, maybe they reach a better place in the Vaporware of the year.
2003 they reach only place 9 with amiga OS4.
http://www.wired.com/science/discoveries/news/2004/01/61935?currentPage=all
but what happen if Hyperions SUper project is Port Duke nukem forever to OS4 ?
http://en.wikipedia.org/wiki/Duke_Nukem_Forever
"""
DNF team fired and 3D Realms downsized - May 2009
"""
Maybe Hyperion port it to OS4 and all that love to play it duke nukem forever now, must buy a SAM.
but if thats possible that Hyperion can port a modern game is the next question.i see not that hypeiron port in the last years a modern game to Mac or Linux.
-
So it'll be a near rewrite for anything AOS based API?
It would be a rewrite for any OS not already designed for many core, I'd say. CUDA (and by extension OpenCL) require that you take a completely different look at how to write code. Essentially you are writing code that launches many (read thousands) of concurrent threads at once over different sections of a dataset that you've divided into a grid. It's like SIMD but a bit more flexible in that it is possible for each thread to take a different path of execution at a conditional branch (you get a penalty when that happens though).
Unfortunately, not all code can be reworked for the many core approach. Only problems that contain inherent parallelism are suitable.
-
You could argue that x86 != AMD64 ;)
That's my thinking too. In fact, a Frieden has expressed his love for AMD64...
-
AMD64 is a pretty nice architecture, once you get past the x86 subset quirks.
-
I still think the APM 83290 is more likely. I posted the link earlier in this thread. With production expected in Q1 2010 and samples available now, this daul-Titan 1.5Ghz chip looks well placed for a NG amiga.
-
And it would play well with the audience, the audience that is likely to purchase AmigaOS anyway, which is all that really counts. It's conservative so it fit's the Frieden's approach. I think you are likely right. Plus it would be a good chance to start the multi-core development.
Multi-core is tricky to get right, and then time for the software to support it. You can't even begin to test if you don't have multi-core machines in your user's hands.
I still think the APM 83290 is more likely. I posted the link earlier in this thread. With production expected in Q1 2010 and samples available now, this daul-Titan 1.5Ghz chip looks well placed for a NG amiga.
-
Nobody exept Hyperion themself knows which direction they are going. But I noticed that multi-core was mentioned on the roadmap for OS4.2.
Turns out there are no plans for 4.2 and SMP is still just a pipe-dream:
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=29845&forum=14&start=120&viewmode=flat&order=0#517864
-
I just read an announcement about Tilera's new 100 core CPU, the Tile-Gx.
Wouldnt porting OS4 to this CPU be a worthy candidate to the speculations around Hyperions secret "most amitious project"?
Well, ever since "ssolie" spilled the beans all over MorphZone.org, and then made the announcement on the AmiWest event, I think it's pretty clear that it's about "the box", and the potential features that might lead to (of which "SMP" (kind of, at least) would be one). And frankly, I couldn't think of a more ambitious project than this.
-
What are you talking about? A search of MorphZone reveals only that ssolie's posts are all about amiga.org getting a new owner and his Amiwest report getting rejected because it had sod all to do with MorphOS. Please, provide some links or your post is meaningless to anyone else.
-
What are you talking about? A search of MorphZone reveals only that ssolie's posts are all about amiga.org getting a new owner and his Amiwest report getting rejected because it had sod all to do with MorphOS. Please, provide some links or your post is meaningless to anyone else.
Well you know, he's hardly what you would call a "regular" at MorphZone, but suddenly one day, he starts posting all over the place about the "box" design of MorphOS, how that was the one and only essential feature of the OS, how that was really what everyone wanted, what we were promised, and how the MorphOS team failed so radically to deliver it, yada yada.
The peculiar thing about it was that it was so *completely unprovoked*, like a flash from a blue summer sky! I mean, nobody has even given the "Q-box" a thought since half a decade or so, but suddenly there he was, pushing this issue like mad. This caused people to wonder why he all the sudden was so interested in box designs, and I know I wasn't the only one who got the feeling that he suddenly "got to know something", and what he *really* wanted to say was: "That initial feature list of MorphOS, with a 'boxed' (microkernel) design, where the 'Amiga' environment runs untouched in a box and new exciting features (like SMP, MP, etc, etc) is added outside it, that feature list was promised to you MorphOS suckers by the MorphOS team a long time ago, this was probably what made you interested in MorphOS instead of OS4 in the first place, and the MorphOS team failed to deliver that to you, but guess what *Hyperion* has up in their sleeve... ... ...". Like he was preparing the ground for the news in his own, weird kind of way.
Here is how he suddenly increases his post count several hundred percent in a very short time...
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6248
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6285
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6287
...and recently this was quite confirmed by the announcement from AmiWest. A box design would be the only even remotely possible way to add "SMP" to the Amiga environment.
-
Thanks for clearing that up, links etc...
It never ceases to amaze me just how much some people get caught up in the whole "this direction, that direction" fanboydom, even within this tiny group of hobyism we all belong to. WOW.
I like to think that having fun is what matters. Sure, I'd love to see AmigaOS on top but I have a bid on ebay for a MacMini right now to run MorphOS and I'll buy a SAM440 as soon as my wallet lets me. Even so, I still have no intention of ever getting rid of my A1200 w/. 3.9, she was my first computer after all. :)
....and yeah, his posts do seem a bit....odd, but hey,I say; Until it's in your hand it's all vapour, right!
-
Well you know, he's hardly what you would call a "regular" at MorphZone, but suddenly one day, he starts posting all over the place about the "box" design of MorphOS, how that was the one and only essential feature of the OS, how that was really what everyone wanted, what we were promised, and how the MorphOS team failed so radically to deliver it, yada yada.
The peculiar thing about it was that it was so *completely unprovoked*, like a flash from a blue summer sky! I mean, nobody has even given the "Q-box" a thought since half a decade or so, but suddenly there he was, pushing this issue like mad. This caused people to wonder why he all the sudden was so interested in box designs, and I know I wasn't the only one who got the feeling that he suddenly "got to know something", and what he *really* wanted to say was: "That initial feature list of MorphOS, with a 'boxed' (microkernel) design, where the 'Amiga' environment runs untouched in a box and new exciting features (like SMP, MP, etc, etc) is added outside it, that feature list was promised to you MorphOS suckers by the MorphOS team a long time ago, this was probably what made you interested in MorphOS instead of OS4 in the first place, and the MorphOS team failed to deliver that to you, but guess what *Hyperion* has up in their sleeve... ... ...". Like he was preparing the ground for the news in his own, weird kind of way.
Here is how he suddenly increases his post count several hundred percent in a very short time...
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6248
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6285
http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=6287
...and recently this was quite confirmed by the announcement from AmiWest. A box design would be the only even remotely possible way to add "SMP" to the Amiga environment.
I wouldn't condone flaming from ssolie are anyone else. However I think it hypercritical coming from you. I'm sure there are plenty of seasoned Amiga forum members will know, that you have persistently engaged in similar behavior over the years and worse.
-
I wouldn't condone flaming from ssolie are anyone else. However I think it hypercritical coming from you. I'm sure there are plenty of seasoned Amiga forum members will know, that you have persistently engaged in similar behavior over the years and worse.
I'm not moralizing upon his posts or posting style over at MorphZone.org, but merely showing how quite a few of us was hinted about the "SMP"/box thing back then already thanks to him. The topic of the thread is speculations about "most ambitious project candidates", I provided a candidate, someone asked for links and an explanation, and I provided that as well.
Your ungrounded attack on me here is both inappropriate and off topic.
-
Why is this whole thing a secret anyway? (Perhaps in case it fails?)
-
Why is this whole thing a secret anyway? (Perhaps in case it fails?)
It's just how communication is done in amiga land, openness is unheard of.
-
Perhaps it's this?
(http://www.baronbob.com/catsharpener.jpg)