@MZ

Quote:

MZ wrote: Compliant cores Cores that comply with the Power ISA v.2.04 POWER6

You're misreading the page. What the page says is that *Power ISA v.2.05* compliant cores consists of all Power ISA v.2.04 compliant cores PLUS POWER6.

If you look at the section above it specifies the Power ISA v.2.04 compliant cores as the Power ISA v.2.03 compliant cores PLUS the PA6T and the Titan. So the Titan (assuming PA Semi are out of the running) seems most likely, though technically it appears older cores are also compliant with v.2.04

_________________
Wiki for new/returning Amiga users - Projects: ACE basic compiler / FrexxEd / Git

@pertinaxone

Quote:

From what I see in this board, it has everything to inspire people from the bedroom coders and hack-it-hobbyists, the demoscene etc all the way to professionals in big companies who can't quite get their Wintel box or Mac to do what THEY want in that slightly different way.

I understand that, but i'm not a coder, i'm an end user who likes Amiga. Won't the hobbyists who is interested in this stuff already have bought the ¤100 XMOS dev kit? Don't get me wrong, I like the sound of the the CPU power and I want one . Hopefully Hyperion are coding something that integrates the XMOS stuff with the CPU and OS, without this then I am not sure I really understand the hype.

@vidarh

Ok... I see.

So 2.03 is still an option.

@pertinaxone

We don't have one but we would LOVE to have two. Developers are important to the success of X1000.

_________________
Sent from my Quantum Computer.

@Radfoo

Quote:

Radfoo wrote: I understand that, but i'm not a coder, i'm an end user who likes Amiga. Won't the hobbyists who is interested in this stuff already have bought the ¤100 XMOS dev kit? Don't get me wrong, I like the sound of the the CPU power and I want one . Hopefully Hyperion are coding something that integrates the XMOS stuff with the CPU and OS, without this then I am not sure I really understand the hype.

The XMOS dev kits are external. That limits what you can do with them significantly. From what has been revealed so far it appears the IO lines on the XCORE in the X1000 can or are bridged to the CPU and other internal components, as well as are available to Xorro cards.

Depending on how tightly integrated it is that could be very exciting, but it'd be weird if it's not far tighter integration than what you can get with the dev kit. In any case, the excitement is likely to come from thinking about how we can use it to enhance the X1000 experience, not specifically the XCORE's themselves. There's already a SID implemention on their website, for example - likely people will come up with all kinds of chipsets to emulate using the XCORE to add capabilities to the X1000 without burdening the main CPU at all.

_________________
Wiki for new/returning Amiga users - Projects: ACE basic compiler / FrexxEd / Git

@MZ

Quote:

MZ wrote: @vidarh Ok... I see. So 2.03 is still an option.

Technically yes, but I don't think they'd say 2.04 if they were going to do that - it'd be kind of underhanded...

_________________
Wiki for new/returning Amiga users - Projects: ACE basic compiler / FrexxEd / Git

@vidarh

Quote:

Technically yes, but I don't think they'd say 2.04 if they were going to do that - it'd be kind of underhanded...

Note that the link X is many things stated:

Quote:

The AmigaOne X generation will come with a variety of CPUs, conforming to the Power ISA 2.04 and newer standards.

Newer standards... wouldn't that mean 2.05, 2.06 and later on 2.07 etc.?

@vidarh

Quote:

Depending on how tightly integrated it is that could be very exciting, but it'd be weird if it's not far tighter integration than what you can get with the dev kit

Yeah, I guess that is the info i'm after. I suppose we have heard the side of the story from A-Eon, we just need the info from Hyperion now to fill in the gaps.

Hope we won't have to wait too long

@Smurfen

I can't imagine they would be able to get a Power7 at an affordable price, but if they did the Amiga would be way ahead of the pack performance wise. Here's to dreaming.

Last edited by TheKorn on 05-Jan-2010 at 11:48 PM.

_________________
Raptor BlackBird (awaiting a Miggy like OS)
Amiga 4000 3.9 / Pegasos II 4.1 F.E.

@TheKorn

If only money was no object,
one can always dream

Johan wrote:

And wouldn't it be a good idea that the well-known jounalist, who thought this all up, wrote a book on how it all happened? Beginning at the very first start, when some-one first started to think about developing such a machine, untill these wonderful and yet tantalizing days at the end of 2009 and the beginning of 2010? The whole story?

@Johan

Great idea... I'd buy that book too...

@KimmoK
I very much doubt that OS4 is running on first core, and Linux on the second core. Plus there has been NO hints in this direction.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@Radfoo Quote:

Also, anyone know if the hidden.png was in reference to XMOS chip or the CPU?

I think it was the XMOS chip: Each thread is only capable of running at 25% of the chip's full capacity, and it has 8 threads, just like shown in the hidden.png.

edit: OTOH, the XMOS chip can only run 8 threads each at 12.5% of full capacity, so there would have to be some odd reporting going on for it to be a single-core XCore chip.

Last edited by ChrisH on 06-Jan-2010 at 12:13 AM.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@Radfoo & ChrisH

Quote:

Also, anyone know if the hidden.png was in reference to XMOS chip or the CPU?

Definitly the CPU

Just look at the picture
http://a-eon.com/gfx/hidden.png

It says:
Quote:

Q: Why are we running 8 tasks at 25% of CPU power each? A: We were trying to get power consumption over 20W

So this literaly says "CPU"
It also says 20W, the XCore would need a clockrate at way over 10GHz to reach 20W... that is just not possible

@Derfs

Quote:

Derfs wrote: you have 2 choices # The PA6T core from P.A. Semi # Titan from AMCC as those are Power ISA v.2.04 compliant

Apparently the Power6 is Power ISA v.2.04 compliant also

Power6 is ISA 2.03 compliant and therefore is 2.04 compliant. It is also specifically listed as 2.05 complaint on the same page, and is therefore by nature, isa 2.06 compliant. Nice.


Quote:

Power ISA v.2.03 The specification for Power ISA v.2.03 is based on the former PowerPC ISA v.2.02 in POWER5+ and the Book E extension of the PowerPC specification. The Book I included five new chapters regarding auxiliary processing units like DSPs and the AltiVec extension. Compliant cores * Freescale Power PC e200, e500 and e700 * IBM PowerPC 405, 440, 460, 970, POWER5 and POWER6 * IBM Cell PPE Power ISA v.2.04 The specification for Power ISA v.2.04 was finalized in June 2007. It is based on Power ISA v.2.03 and includes changes primarily to the Book III-S part regarding virtualization, hypervisor functionality, logical partitioning and virtual page handling. Compliant cores * Cores that comply with the Power ISA v.2.03 * The PA6T core from P.A. Semi * Titan from AMCC Power ISA v.2.05 The specification for Power ISA v.2.05 was released in December 2007. It is based on Power ISA v.2.04 and includes changes primarily to Book I and Book III-S, including significant enhancements such as decimal arithmetic (Category: Decimal Floating-Point in Book I) and server hypervisor improvements. Compliant cores * Cores that comply with the Power ISA v.2.04 * POWER6 Power ISA v.2.06 The specification for Power ISA v.2.06 was released in February 2009. It is based on Power ISA v.2.05 and includes extensions for the POWER7 processor and e500-mc core. One significant new feature is vector-scalar floating-point instructions. Book III-E also includes significant enhancement for the embedded specification regarding hypervisor and virtualisation on single and multi core implementations. Compliant cores * Cores that comply with the Power ISA v.2.05 * POWER7 * e500-mc core

It'd be nice to have confirmation from someone in the know as whether the CPU is 64-bit or not, as Titan seems to be only 32-bit.

Last edited by tiffers on 06-Jan-2010 at 01:39 AM.
Last edited by tiffers on 06-Jan-2010 at 01:34 AM.

@EntilZha

Quote:

EntilZha wrote: @maz77 Quote: Do A-Eon have plans to market the A-one X1000 with Linux and OS4 together? I don't know what plans A-EON has regarding distribution of the machine. I'm just the kernel guy, you know...

But do you know if the Kernel would, say, use both processors of a dual core CPU?

_________________
Rage quited 29th May 2011

@Cameraman

It happened, but only in German: "Der Amiga, die Geschichte einer Computerlegende" by Volker Mohr (that's me ). You can buy it at Amazon, Vesalia.de, or in any book store. I really would like to make an update (it came out in 2007), and if somebody likes to translate it (a native speaker, if possible) and has contacts to a publisher, feel free to contact me. The problem before was, that my publisher, "Scriptorium Verlag" has no distribution network or contacts in the UK/USA.

@EntilZha

I'd just like to say how much I enjoyed the mystery over the last few days, a whole lot of work went into setting those puzzles and putting it all together - very nicely done AK. And yes, much better than just splurging out a cold press release.

The X1000 board looks to be very well designed and the spec is great. Congratulations to everyone involved to bring the project to this stage. Best wishes for the New Year, and the continued development of this MAP. I hope there will be much success with the X1000.

My money will be waiting, that's for sure
Aproximate cost info would be good to know though

Last edited by raumfahre on 06-Jan-2010 at 03:41 PM.

Well, finally finished catching up with what's been posted here and at other sites. It's been a fun ride, staying up till the wee hours and wildly speculating the tiniest bit of info on the a-eon site. Can't remember the last time I had that much fun with a bunch of guys over the web, and much of what I've read has made me think. Time to take a step back, and ask "what have we actually got here"?

1. An Amiga computer that's up to date. About frickin' time I say.
2. Strong hints that AmigaOS has finally gone multi core. They said it couldn't be done, and it looks like they were wrong.
3. A thread over two thousand posts long about OS4 related material that has only a few isolated instances of trolling - is the community finally putting its idiot years behind it?

Then there's Xena. To me, the Xena chip is a huge question mark. Yes, it's a very cool coprocessor. The thing is, it doesn't "fit" with the rest of the design. Its strongest point is that it's deterministic (and that's *huge*), but as far as I can determine it's only connected to the rest of the system via the PCIe bus. PCIe is not a deterministic bus (other devices can take over at their discretion at any time). That means that you can only use Xena's determinism if you connect external hardware to the X1000. Let me return to that in a moment. The other salient points of Xena are that it's capable of 8 threads of execution and has 400 MIPs worth of oomph. The PA6T-1682M (one of the contenders for the X1000 CPU) running at 2GHz (the rumored speed of the CPU) clocks in at 8,800 MIPS (source). That means that if Xena wasn't present, the main CPU would be able to to pick up the slack using only ~5% of itself. AmigaOS is a very efficient OS, so the task switching overhead for the 8 threads would be negligible. So Xena doesn't really add anything to the overall computational ability of the X1000. Yes, one can add more XMOS chips via the Xorro slot, but this would lead to "to run this game, you need the Individual Computers Quad Xena and clockport card". Unless this card is bundled with new X1000s, I don't see that program selling too many copies - think back to the days of A500s and FPUs: how many programs required an FPU at the time? So Xena's greatest strength (determinism) doesn't add anything to the system and its processing capability only adds about 5% extra cycles. That's why I think it doesn't "fit". But the guys who put it there must have had a good reason, and I don't think it's just a marketing gimmick or nostalgic appeal. Why then? Simple: look at its little brother, the SAM440. It's got an FPGA on board, which as far as I can tell is still pretty much unused by the community (I'm sure there are some exceptions). BUT! it has a very important purpose: it makes the board attractive to industrial customers, who subsidize the board for the rest of us. Can Xena be the same thing? I think so: it's pretty much useless for the average amigan, but it gives the X1000 something that is very desirable to industrial and research customers: hard time realtime capability in a desktop class computer (when you plug in the extra hardware these folks use). Xena's job is to make the board affordable by finally allowing us to take advantage of economies of scale.

Last edited by CodeSmith on 06-Jan-2010 at 10:12 AM.

@ChrisH

"I very much doubt that OS4 is running on first core, and Linux on the second core. Plus there has been NO hints in this direction."

If they do not use the possibility, I will be surpriced.
It would be an advantage over any competing product.
It would also enable AOS4 on the HW without the HUGE SMP work, while being able to use both cores.
It would deliver a kind of memory protection. Apps would have two separate memory spaces.

If not Linux + AOS4, then AOS4beta + AOS4final setup.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?