> I am not sure what caused the design delays, or what mistakes were made, if any

I think one of the biggest pre-2010 changes in the X1000/Nemo design was the scrapping of several dedicated I/O chips in favour of the SB600 southbridge chip:

"the prototype Nemo motherboard development took much longer that originally estimated and much of the delay can be attributed to a change to the original design specification by the inclusion of an SB600 Southbridge, which although a technically superior solution, created several additional challenges in the manufacture and testing process."
http://safir.amigaos.se/main.php?d=newsfull&news=1133

"We had some difficulties with the getting life signs out of the SB600 (South Bridge)"
http://obligement.free.fr/articles_traduction/itwgentle_en.php

"The customer chose a PC motherboard South Bridge to handle the peripherals, and presented us with the challenge of getting it working within a PowerPC architecture CPU."
http://www.amigans.net/modules/xforum/viewtopic.php?post_id=51602#forumpost51602

"The SB600 doesn't follow normal PCIe address decoding rules, as in its original use as an AMD south bridge it was designed to appear as if it were integrated into the north bridge."
https://lists.ozlabs.org/pipermail/linuxppc-dev/2016-June/144114.html

The 2010+ delays that followed were more delays on top of that.


Edit: added another quote

[ Edited by Andreas_Wolf 03.08.2016 - 01:18 ]

Quote:

And on request (from the thread mentioned above), here are some Blender results:

http://www.amiga.org/forums/showthread.php?t=61508



[ Edited by takemehomegrandma 04.05.2012 - 20:23 ]

To cut the blahblah short, I tried to do summary with blender versions that are in the same ballpark:

With blender 2.49 or 2.48 & dualcore:
120s blender 2.49 G5 2.3Ghz Dualcore
121s blender 2.49 G5 2Ghz Dualcore (erroneously reported singlecore?)
127s blender 2.49 G5 3Ghz Dualcore (erroneously reported singlecore?)
206s blender 2.48? x1000 1800Mhz Dualcore

Then with blender 2.49 or 2.48 & singlecore:
223s blender 2.49 G5 2.3Ghz single core
234s blender 2.49 G5 2.1Ghz single core
235s blender 2.49 G5 3Ghz SingleCore
325s blender 2.49 PowerBook 1667Mhz
331s blender 2.49 PowerBook 1667Mhz
420s blender 2.48? x1000 1800Mhz Single core
425s blender 2.48a PowerBook 1500Mhz
540s blender 2.48 MacMini 1420Mhz
591s blender 2.48 A1XE 1000Mhz
1180s blender 2.48 SAM440 800Mhz (AOS4)

With same blender version G4 is slower than G5, but faster than PA6T.

Performance scaling etc.:
G5-2300single vs G4-1667 - 1.3 x Mhz - 1.4 x result - G5 better

G5-2300single vs PA6T-1800single - 1.28 x Mhz - 1.88 x result - G5 better

G4-1667 vs PA6T-1800single - 1.08 x Mhz - 0.77 x result - G4 better
G4-1667 vs PA6T-1800dual - 1.08 x Mhz - 1.6 x result - PA6T is better

The last part shows PA6T advantage, one day, after AOS is updated to use both cores. (it seems G5 does not scale as well with multiple cores, neither should G4)
(not that any of this would make x1000 pricing more sensible towards end user)
For blender, what ever PPC one uses on the designer/modeler, one would/should consider x86 render farm for speedup, instead of G5 or PA6T machine, IMO.

[ Edited by KimmoK 07.05.2012 - 11:29 ]

Update:

>> The new website should be up and running soon.

> Also on http://www.a-eontechnology.com? ;-)
> http://www.networksolutions.com/whois-search/a-eontechnology.com

...has now been changed to point to the Hyperion message boards instead of the old A-Eon website.

Reminder:

> Hyperionmp says:
> "The way in which it will be implemented however is already clearly defined
> and was subject to peer review by other developers. Obstacles to an
> efficient implementation were removed (e.g. the use of Forbid) and replaced
> in many OS components over the years (e.g. DOS). The foundation for
> SMP support was put in place, a clear picture exists what needs to be done
> to accomplish it and how. I'm willing to take a bet that it won't take 2 years ;)"
> http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=34171&forum=33&start=40#627520

One third has passed already :-)

@Andreas_Wolf

Thought you might be interested in reading this:

XENA - Signs of life

#6

> Thought you might be interested in reading this

I am. Thanks for pointing me to it. Now that OS4 can access the XCore chip there remains "only" the XMOS SDK to be ported to OS4.

Sending this message from my X1000 computer wirelessly connected to my WLAN via the NetGear WNCE2001-100NAR that I purchased from NewEgg as a refurbished product priced at $19.99 with free shipping, but I also had a $10 off coupon from a previous purchase, so my total price including all taxes and the free shipping was less than $12 delivered to my front door this afternoon.

I plan on testing it with my G4 PowerBook running MorphOS3.0 beta in a few minutes and I am sure it will work just as easily.

I really like the small size of the NetGear WNCE2001-100NAR device and the fact that it can run from the included USB power cord, or the included PSU. The next product to search for is a battery powered device that has a USB port that I can plug the USB power cord into so it will give me a temporary completely wireless solution for wireless Internet Browsing using this device with my G4 PowerBook, while I am waiting for the NetGear WG511T wireless CardBus NIC to be supported by MorphOS3.x some time in the future. Maybe the USB port on the G4 PowerBook can provide just enough current to power up this NetGear WNCE2001-100NAR, even if they don't provide enough electrical power for external hard drives.

Another possibility is perhaps someone makes a FireWire400 to USB port adapter, just for providing power to this device?

I will edit this message in a few minutes from my PowerBook if I can get it setup and working with the same NetGear WNCE2001-100NAR.

Thanks to who ever it was who suggested this device as a possible solution to wireless connections to my WLAN.

Quote:

Finally an area where the A1X1K is winning over the 1/20th priced competition - Flashing LED's!

Update:

> As for the maintenance statement, maybe this refers to the
> March 2010 announcement of a more professional website:
>
> "we will have a more traditional corporate website at the
> A-EON address when we are ready to go fully public."
> http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30985&forum=42#544604

It's there now.

New site is absolutely disgusting, Aeon guys should take some example from morphos.net site :)

Somehow they managed to squeeze in 99 errors and 7 warnings in that rather small web page:

http://validator.w3.org/check?uri=http%3A%2F%2Fa-eon.com%2F&charset=%28detect+automatically%29&doctype=Inline&group=0

That's actually kind of an achievement, using only about 100 lines of actual HTML-code (embedded CSS and JS removed).

It's using tables for layout/graphical design, which is a death sin! That was common back in the Ibrowse days, when there was no CSS, but you simply don't do that today. And I am puzzled about why, since the page *also* uses CSS, which is the proper way of doing layout/graphical design.

Speaking of which, CSS has 40 errors and 46 warnings:

http://jigsaw.w3.org/css-validator/validator?uri=http%3A%2F%2Fa-eon.com%2F&profile=css3&usermedium=all&warning=1

IMO, the looks of the page suggests it was made by someone learning HTML, the validation kind of confirms it. But I guess it's at least an improvement from the old web site, with that dildo shaped menu penetrating the page from the lower right...?

@amigadave
Is MorphOS 3.0 out there yet , or are you a betatester?

> Is MorphOS 3.0 out there yet , or are you a betatester?

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7967&start=48

@ThEcRoW,

No, I am not a beta tester per se, but I was given access to an early version of MorphOS3.0 to show at last year's AmiWest 2011 Show.

I had some difficulty with connecting to the hotel's Internet connection using an Ethernet cable, which had a large negative effect in my preparations for my presentation during the show, so I was very disappointed (again) with the presentation that I gave.

I am hoping that MorphOS3.0 will be finished and released before this year's AmiWest 2012 Show, so I can have a second chance at giving a better presentation (unless someone else wants to volunteer to do a MorphOS presentation at this year's show).

My attendance at this year's AmiWest 2012 Show is dependent on my health and a possible second back surgery that may occur some time this year.

I am not a beta tester, as I would assume that they work with the MorphOS Dev. Team in reporting bugs and giving other reports on performance, or other such things. I do not do that and actually have almost zero contact with any of the MorphOS Dev. Team members and what little contact I do have with any of them, does not mean that I am getting any information about MorphOS3.0 that the rest of you don't already know. I don't ask and they don't tell me anything.

Andreas_Wolf,
Quote:
I used that at my university! XC is pretty frustrating to code in. Concurrent C with loads of deadlock.

I've met Ali Dixon from XMOS too - he was happy to refer to the X1000 as "vapourware". :)

[ Edited by Mequa 03.06.2012 - 22:18 ]

Mequa,
Quote:

I guess Ali Dixon was wrong, wasn't he? My "vaporware" X1000 is sitting on my desk in the other room with the XMOS chip inside of it on the Nemo2 Motherboard.

What do you mean by "Concurrent C with loads of deadlock"?

> I've met Ali Dixon from XMOS too - he was happy to refer to the X1000 as "vapourware". :)

I remember you've told about that incident already elsewhere in December 2011 :-) And as events have turned out some weeks later, he was wrong.

amigadave,
Quote:
XC programming language - I assume that's the primary language used to program Xena?

We were taught that along with CSP. The latter is used for formal reasoning about concurrent programs. That's pretty essential when getting multiple XMOS cores or threads to communicate, or the XC program will have all kinds of race conditions.

Most students on my course struggled with the shift in paradigm from sequential programming, it really requires a lot of formal planning to avoid deadlock. Even for seasoned programmers. What didn't help was the bugs in the version of XC we were using - two semantically-equivalent programs gave different results. Hopefully XMOS will debug their SDK.

Quote:

At $3000+, nobody (other than a few... ehrm, "brave" ones) would ever buy something like it today, even if they could. Which they can't, since it's not up for sale. There has been 2 (or three?) small batches produced, on a pre-order scheme (and probably pre-payment as well), with sporadic deliveries many months later. There is very poor SW support for it, especially in OS4 which is the whole point of the HW. It relies entirely on a CPU that never reached a commercial phase, samples are limited (and the company behind it doesn't exist anymore, the design is as dead as it can be), and so is the potential production volumes of the motherboard in question, even if it would have customers, which it hasn't, since it's not out for sale, and wouldn't have had even if it were, since they are asking more than $3,000 for this piece of 2005 specced HW. So for anyone developing real products, for real markets (like XMOS does), this is just as much a *none-product* as vapor is, and this won't change!

[ Edited by takemehomegrandma 04.06.2012 - 10:49 ]

amigadave,
Quote:

Hmmm.. that's in late October isn't it?

I admire your optimism

Quote:

>>> I've met Ali Dixon from XMOS too - he was happy to refer to the
>>> X1000 as "vapourware". :)

>> I remember you've told about that incident already elsewhere in
>> December 2011 :-) And as events have turned out some weeks later,
>> he was wrong.

> At $3000+, [...] a few... ehrm, "brave" ones [...] would [...] buy
> something like it today [...].

Exactly that's what happened.

> it's not up for sale.

True, it's not up for sale *right now*, but it was up for sale in the past and there is sufficient proof that it has been purchased by people back then. The term "vapourware" does not mean something is not up for sale right now, but does mean that something has never been up for sale (and probably will never be).

> There has been 2 (or three?) small batches produced

To date it's been just one single batch ("First Contact") sold to regular customers:

http://www.a-eon.com/news/25-01-2012.html

> probably pre-payment as well

Yes, this is a known and admitted fact:

"Will the money be taken stright away or only at shipping? - AmigaKit will take a deposit, final payment will be required before delivery."
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=34475&forum=33&start=20#634850

> It relies entirely on a CPU that never reached a commercial phase

AFAIK, the PA6T-1682M went from sample to production level somewhen between late 2007 and early 2008. Whether the PA6T chips in possession of A-Eon are sample or production level chips I don't know. If my timeline of events resembles the truth it should be production level ones.

> so is the potential production volumes of the motherboard in question

True, the number of Nemo boards that can be produced for the X1000 is limited by the number of PA6T chips they have in possession. I don't know what that number is and so don't you. Of course we can speculate about that number, but that won't affect my opinion about the X1000's (non-)vapourware status.

> even if it would have customers, which it hasn't [...].
> and wouldn't have had even if it were, since they are asking
> more than $3,000

Some sentences before you stated that is has "a few... ehrm, "brave"" customers paying that price. So which one is it?

> since it's not out for sale

It was out for sale, and the people who purchased it are existing customers, according to my definition of that term at least. Even if there was no further batch to ever be sold to new customers, this wouldn't make the already existing customers non-existent. There's even someone attempting to put up a (non-exclusive) list of existing X1000 customers:

http://moobunny.dreamhosters.com/cgi/mbmessage.pl/amiga/217870.shtml

> this piece of 2005 specced HW.

I think it's more like 2007-specced hardware. The graphics card it comes with was released in 2008 btw.

> for anyone developing real products, for real markets (like XMOS
> does), this is just as much a *none-product* as vapor is

I don't share your definition of the term "vapourware", and thus I still think Ali Dixon was proved wrong. The difference between him and you is that he made his "vapourware" statement prior to the X1000 being put up for sale to regular customers whereas you keep on making it even now.

@Andreas

It's simple really. You can't buy it, there aren't any available, it's not "in production", and except from one or two limited batches (with very special circumstances attached to it, see below), it has *never* been. One of the key parts of the equation, the SW part (OS4, the very base for the "AmigaOne" name license, for the boing ball on the case, the single one reason to *the very existence* of the AmigaOne X1000 concept in the first place), is even quite far from supporting it as it should, it will possibly take several years still until it does, if ever. They have made limited production batches, the first one sold as "Beta tester" boards under NDA. Sure, our views on the term "Vapor" may differ, but IMHO when something is being made in extremely limited batches, without key components being finished yet, and the thing is labeled "First Contact", "Early Bird", whatever, and when customers are asked not to talk about their experiences in public but in closed, private forums, etc, when customers are discouraged from testing/benchmarking the thing, etc *with references to its unfinished state*, then I think it's more appropriate to call it "Prototype" being sold while in development, rather than "Product", and IMHO this is what a very limited bunch of people are in possession of. Prototypes. Not a finished product. There are some people in possession of the Amiga Walker from Amiga Technologies. There are some people in possession of the C65 from Commodore. There are many more examples of things actually existing in a tangible form (read: the results of something coming out from a very limited production run), but prototypes nevertheless, that has not yet reached its full specifications. The AmigaOne X1000 as a whole has not yet reached its 1.0 level. It has never been sold from stock without prior written agreements between the customer and the producer directly, like normal products would. I couldn't order one this instant and have it shipped to me today, even if I wanted to, and this has *never* been possible! Never! It's not here, it's not even finished. It's vapor. Prototype maybe, but it's a none-product nevertheless.


[ Edited by takemehomegrandma 04.06.2012 - 14:20 ]

Speaking of the A1X1K, someone over at moobunny seems to have compiled a list of known owners:

http://moobunny.dreamhosters.com/cgi/mbmessage.pl/amiga/217870.shtml

Quote:

Does anyone know about a nick/name of someone that is missing from the list, that should be added there?

[ Edited by takemehomegrandma 02.07.2012 - 17:27 ]