New Hyperion Entertainment Website http://a-eon.com/ - The Mystery Continues

[takemehomegrandma]

I'm guessing there is a 3rd way, which is what MorphOS finally did in supporting 2nd hand PPC Macs.

I agree!

You can get HW for $50-$350 depending on what you need and want. Support for Mac Mini is here today, support for PowerMac is close, and after that follows (probably) the PowerBook support. Desktop to Laptop, via the Mini.

A Mac Mini 1.4 GHz is about 2x as fast as a Pegasos2 in some areas. And you can get it from everywhere.

A little late IMO, but better now than never.

They obviously focused on getting a full MorphOS 2.0 out. Then they had to deliver the Efika 5200B port (which they could have done much earlier, based on some 1.4.x version, but this is how it was played out).

This is not a solution for world domination, but give retro hobbyist something nice to play with at a dirt cheap price. Hyperion taking this route with OS4 would have made more sense to me.

If either the MorphOS team or Hyperion really want to take a run at mainstream, I don't see any other option than x86.

I agree:

http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=11&topic_id=6684&sortname=&sortorder=&sortdays=&viewmode=flat&order=0&start=171

[hazydave]

http://www.power.org/devcon/07/Session_Downloads/PADC07_Chang_AMCC_Titan_V0_2.pdf
    * Cores that comply with the Power ISA v.2.03
    * The PA6T core from P.A. Semi
    * Titan from AMCC

Titan AMCC anybody? -

[/quote]

I think that's a good guess. For one, it just started shipping at 1.5GHz last October... they say you probably haven't seen the chip yet (true), and something about clocking at 1.6GHz (implying that's the not the normal speed).. regular speed is 1.5GHz. And all the SOC I/O matches the stuff they're claiming for the main board. Here's the AMCC paper on this from the 2007 PowerPC Devcon:
http://www.power.org/devcon/07/Session_Downloads/PADC07_Chang_AMCC_Titan_V0_2.pdf

Manufacturered on the cheaper 90nm CMOS process and other cost reductions
Design allows for dual cores at under 15watts power consumption

They's using some weird NMOS with mult-phase clock (6502, anyone?) and dynamic logic, like CPUs from back in the 80s and early 1990s, to cut down on speed demands. Certainly better than the stuff that's been out in the Amiga market, but a shame they couldn't go to a real [at least formerly] destop-class CPU like the PPC970.  The AMCC has an FPU, but no vector unit. So, multimedia performance is even worst than it might have been with a modern applications processor.

And of course, they would be crazy to use a PA Semi chip... scraps from Apple's table that could be withdrawn at any minute.

Only thing.. that huge heat sink area show on the MB is crazy overkill for the AMCC chips.

They claim 4000 Dhrystone MIPS per core for the AMCC at 2.0GHz, so that's 3000MIPS per core at 1.5GHz. That's in the range of a Pentium 4 or Athlon/Athlon "Barton" CPU, PC-wise, at least from the one benchmark.. but of course, dual core... great. If they have AmigaOS supporting dual core, of course. Not modern desktop, but let's see... my Q9550 system replaced an Athlon64x2, which replaced an Athlon Barton system... so that's only three geenrations behind the PC state of the art (given my Q9550 is already one generation behind). This is about 1/2 the performance of a Core2 core, or 1/3 the performance of an i7 core, give or take. Of course, it's the FPU and Vector instructions that come to play on the heavy duty multimedia stuff once associated with the Amiga.

Since the Titan just started shipping limited quantities in October of 2009, is dual-core, has a very low thermal design point - the clock speed could be at 1.6Ghz but the cores are rated to 2Ghz currently. Sounds like a contender to me. ;^)

According to the article I saw, it's actually the 1.5GHz version that's shipping. They need a shrink to 65nm to support 2.0GHz+.. at least according to the article.

a-eon site - "Just as Commodore did with the A1000, we're aiming at the high-end first, with a powerful desktop computer aimed at the professional and serious hobbyist markets (although you won't have to wait until summer, and it should be a little cheaper!)"

A professional doing ... what? Does anyone make a living doing AmigaOS development? That's about the only profession I can see happy here.

Commodore Amiga 1000   Price:   US $1295 without monitor   (So under $1295 for those asking about price.)

Yup... back then, Macs and PCs were averaging around $3,000. In 2009, the average desktop computer bought for home use was just over $500 in the USA, and came with monitor and printer.

[hazydave]

Now for the serious question: Does anybody have experience with the XMOS chip? It looks very nice, but what it is capable of?
I undertand that as it uses a parallel architecture, I should not take the 400Mips speed as I would do for an ARM fo example, but is it fast enough for data processing, or it is just something I could use for minor tasks?
Would it make the difference or would it better to rely on raw CPU power instead?

Well, it's a weird ass chip, that's for sure... I haven't used it. Not sure anyone has... they don't explain very well why you would want to, rather than, say, a normal ARM. Their idea seems to be to support a bunch of hardware threads per chip. Now, of course, eight threads on a single core, all running, means each thread is lucky to get 1/8 of the total performance... there's enough delay and waiting in a typical modern CPU to make two threads per core sometimes a good idea (Intel "Hyprethreading"), but eight?

But they also say "event-driven computing". So maybe as an I/O processor? Most threads would be sitting idle most of the time, but spring to life when triggered a hardware event? Remedial I/O processing for folks who haven't discovered multitasking and interrupts, or something more profound? I really do wonder if they have an actual use for this, or just dropped it in to fool Amiga fans into thinking "ooooh... magic chip".

I'd much rather have seen an FPGA that could be addressed on the PCIe bus, as a standard feature. A decent one could be programmed to accelerate all sorts of things that the CPU may be too anemic to do to a modern par. I don't see how a PEAK of 1600MIPS (all four cores and all eight threads per keeping busy) is a big help, even to a 6000MIPS or so host processor. But if it's doing something significant to handle I/O, then maybe it's a good idea. Again, have to see just what they're doing with it.

[hazydave]

IIRC Apple switched to Intel because Motorola were unable to make a G5 processor run cool enough to use in a laptop. This was a serious issue for Apple who were stuck with the increasingly obsolete G4 chips for laptops, they had to do something.

That was a symptom of the greater problem... the PPC just could not compete. Period. The PPC970 caught up to the Operton's of the day (they were a pretty decent match), but that, after how many years. And the big battle between AMD and Intel was still hot. Today's desktop chips are better than twice as fast, in not that many years.

And that PPC970 was just one chip. There were dozens of parts, from many companies, competing in the PC market, optimized for desktop, laptop, server, etc. PPC had and has far better representation for embedded computing, but for PC-class, it was never that strong. This was the inevitable result of Apple's ending of the Mac Clone... the market never got large enough to compare to the x86 market.

Apple was also losing customers and applications. Their market was too small for custom coding and testing, being the only PPC folks around. And there was a certain risk going to a Mac, particularly given the apparent dead-end of the PowerPC on the Mac. And they were charging quite a bit more for quite a bit less CPU.

All of these things lead to Apple switching to the x86. Making the Mac into just a PC in an Apple case, they actually improved the breed. It didn't cost Apple crazy money to keep developing motherboards and system chips.. they could just order them up from Foxcon or whomever.. so they could fall in price... one big barrier to entry. You could run Windows, native, full-speed, etc. as long as Apple supported some way of booting though a standard BIOS... there goes the risk, and that's one big barrier to entry. They follow the CPU, chipset, and peripheral market driven by pretty much everyone except Apple. All very good things for Apple, and their customers.


[/QUOTE]

[hazydave]

I work in graphics... I can tell you. This is never going to get professional markets. If they win a military or government contract for some specialized task then great, but those machines are built to "military spec" (where's Doomy when you need him)

Actually, a very great deal of what the military buys is COTS (Consumer off-the-shelf) or "hardened COTS". The robotics controllers I built for a number of different robots found their way to Iraq in about 3000 robots. Consumer technology, but "hardened" by using higher spec parts where possible. But a far, far cry from MIL-spec. There's no faster way to turn a $5,000 robot into a $50,000+ robot... well, you all have probably heard about what it does to hammers and toilet seats. Same deal.

But for PCs, unless there's some very special purpose, they have all kinds of rules about security, operating systems. etc. And it's usually very difficult to get your company approved by the government for military or other government procurement. My company did, mainly because we had the only R/C controller technology at the time that did what they wanted.

[hazydave]

All this talk of x86 this...x86 that....

Has it ever occurred to anyone the x86 viability roadmap is coming to an end, even with multiple cores?

Let's see.. there hasn't been a new PowerPC desktop chip in six years, but you really think the x86 is ending its life. Despite the fact Intel and AMD sell over half a billion CPUs a year? Despite the fact that Intel is nearly 15% of the IC industry's sales, but nearly half of the profits for the ENTIRE chip industry (hint: the AVERAGE CPU is sold for $6.00.... ?

And despite widely published roadmaps to the contrary?
http://www.firingsquad.com/hardware/intel_32nm_westmere_roadmap/
http://www.firingsquad.com/hardware/amd_cpu_roadmap_update_2008/
http://www.tomshardware.com/news/amd-deneb-heka-propus,6364.html

[AeroMan]

I'd much rather have seen an FPGA that could be addressed on the PCIe bus, as a standard feature. A decent one could be programmed to accelerate all sorts of things that the CPU may be too anemic to do to a modern par. I don't see how a PEAK of 1600MIPS (all four cores and all eight threads per keeping busy) is a big help, even to a 6000MIPS or so host processor. But if it's doing something significant to handle I/O, then maybe it's a good idea. Again, have to see just what they're doing with it.

I have the same opinion. A FPGA could be programmed to emulate AGA for compatibility and other nice stuff. They selected a 400MIPS chip, so based on speed, it is not a big help regarding processing power
If the idea is to use the XMOS to handle some kind of internal IO operations, it is a good one. But if they are aiming at using this chip to control some kind of user port, few people would take advantage of it. Probably, just geeky engineers like me
At some point, they stated that it could be used to emulate SID. I believe most of us would expect this chip to be capable of emulating the Amiga chipset, or doing something brand new. (I bet on the last one)
Anyway,  the XMOS chip seems very nice, as a microcontroller. I could not resist to show that to the folks at work.
(PS: Oh my god, I'm talking to "the man". It's such an honor)

[itix]

@hazydave

I'd much rather have seen an FPGA that could be addressed on the PCIe bus, as a standard feature. A decent one could be programmed to accelerate all sorts of things that the CPU may be too anemic to do to a modern par. I don't see how a PEAK of 1600MIPS (all four cores and all eight threads per keeping busy) is a big help, even to a 6000MIPS or so host processor. But if it's doing something significant to handle I/O, then maybe it's a good idea. Again, have to see just what they're doing with it.

There was an FPGA on Acube's SAM board yet nobody found any use for that. Of course the customer base is very small and there are so few with skills for FPGA programming. Not that XMOS makes difference either but it can be programmed in C at least...

@AeroMan

At some point, they stated that it could be used to emulate SID. I believe most of us would expect this chip to be capable of emulating the Amiga chipset, or doing something brand new. (I bet on the last one)
Anyway, the XMOS chip seems very nice, as a microcontroller. I could not resist to show that to the folks at work.

It could be used to emulate SID chip (to replace entire player there isnt enough SRAM) but making such chip which works as good as original SID chip is such achievement that I doubt it...

[persia]

Why does everyone talk about Motorola PPC?  The original AIM (Apple IBM Motorola) created the chip, it was based on IBM power technology.  Motorola spun off it's IC business to Freescale and sold some power architecture to AMCC.  The current Power.org companies are IBM, Freescale and AMCC...

But that aside, it's pretty much dead technology, used in Game consoles and Micro-controllers, and developed for those industries.  If a big player came along, the Power architecture could be developed to compete with Intel, but if Apple, who was part of the alliance that created it, wasn't big enough to force that kind of development, who would be?

[Crom00]

HazyDAVE...

My comment  about military spec came from a young lady I know who is a project manager on the Hubble Space telescope. She told me that one of the Computers in the thing is a 1990's 486... not a chip from another supplier with equivalent CPU power... it's a 486..they had to go through so much engineering BS to get it approved and fabbed for aerospace use.  It's built to withstand the rigors of liftoff and whatever else happens up there.

Reagardless of all that am I crazy to believe that a FAST Amiga legacy compatible machine can be made using FPGA technology and come in under $300?

I mean go to deal extreme and you see Super Nintendo, and Sega Genesis TV game units for like $30 bucks. Those consoles were more or less Amiga 500 like tech...

I think 99% of us here would pay $200-$250 for a NATAMI like device with an expansion bus for a faster CPU a la the A1200,  or Walker styled motherboard design.

AS for the potential high price.... This is happening because you can bet Acube and Hyperion  watch CVPPC cards go for $800 or more on EBAY and wonder well... if they're willing to pay that much for a 10 year old CPU card they can pay up to 2X that for a new system.

[Hell Labs]

To be honset I want to know WHY the old amigas are so expensive. Every one of it's peers is £10-30 on ebay or the like, yet we pay around £100 for one somebody kept the box for. madness. Though I guess the 1200 and 4000 staying current all the way up to 2004? though I guess a with late 90s early 2000s amiga setup, "1200" is more of a vague description than a model number.

[LoadWB]

Let's see.. there hasn't been a new PowerPC desktop chip in six years, but you really think the x86 is ending its life. Despite the fact Intel and AMD sell over half a billion CPUs a year? Despite the fact that Intel is nearly 15% of the IC industry's sales, but nearly half of the profits for the ENTIRE chip industry (hint: the AVERAGE CPU is sold for $6.00.... ?

I would argue that, technically, x86 is dead.  It exists in RISC emulation layers and x64 is steadily erasing its footprint.

And thankfully at that.  Intel has been flogging that damnedable 32-bit stuff since 1983, meanwhile 64-bit chips have been floating about since around 1990.  It is about time that 64-bit made a strong landing on the desktop.

And as much as people hate Microsoft, it has done something right with Windows 7: in order to obtain WHQL status for a device, a manufacturer must provide 64-bit as well as 32-bit drivers.

[WotTheFook]

I think all A-Eon are selling are rose-tinted Amiga glasses, nothing more; it's like beer goggle surgery for nerds and the number of people seduced by this hype has amazed me.

[LoadWB]

I think all A-Eon are selling are rose-tinted Amiga glasses, nothing more; it's like beer goggle surgery for nerds and the number of people seduced by this hype has amazed me.

And I suppose there is nothing for which you harbor (or is that "harbour"?) a passion, other than trolling?

[Karlos]

I would argue that, technically, x86 is dead.  It exists in RISC emulation layers and x64 is steadily erasing its footprint.

And thankfully at that.  Intel has been flogging that damnedable 32-bit stuff since 1983, meanwhile 64-bit chips have been floating about since around 1990.  It is about time that 64-bit made a strong landing on the desktop.

And as much as people hate Microsoft, it has done something right with Windows 7: in order to obtain WHQL status for a device, a manufacturer must provide 64-bit as well as 32-bit drivers.

That's not really true. The AMD64 architecture is a superset of the x86 one. Of all the x86 features, only 16-bit mode is unsupported in x64 native mode, all the 32-bit x86 stuff is very much part of the architecture still and most of it works when the CPU is running in 64-bit mode too. Hence, you could make the argument that in fact, x86 is continuing to grow.

[runequester]

Let's see.. there hasn't been a new PowerPC desktop chip in six years, but you really think the x86 is ending its life. Despite the fact Intel and AMD sell over half a billion CPUs a year? Despite the fact that Intel is nearly 15% of the IC industry's sales, but nearly half of the profits for the ENTIRE chip industry (hint: the AVERAGE CPU is sold for $6.00.... ?

And despite widely published roadmaps to the contrary?
http://www.firingsquad.com/hardware/intel_32nm_westmere_roadmap/
http://www.firingsquad.com/hardware/amd_cpu_roadmap_update_2008/
http://www.tomshardware.com/news/amd-deneb-heka-propus,6364.html

I have a question to you lot in general, since I don't know a ton about processors.

Its pretty obvious that x86 is ruling the desktop, and it seems ARM has embedded devices as their big thing. Why did the next gen consoles use powerPC processors?

Im not trying to pick a fight or anything, Im just curious. Do they do better in a smaller (physical) machine or less heat or what made them go that way ?