@itix

I think the problem is comparing "classic" with "NG".

It's like a 4 year old child plays perfect Eric Clapton song
and post it on youtube and everyone will agree it's cool.

Then his 40 years old teacher grabs guitar and says "Look!
I'm so much better than him".

The other problem I have is that I can't seem to use more than
one computer at a time so my Amiga 1200 and WinUAE have
cloned configurations with 4 Mb ram so I can test some new
68k thing and burn to CD to transport to Amiga.Under the Amiga
1200 table is small AROS machine now and I'm typing this on
third Windows based computer :P

If I added more systems I wouldn't use any ...

@itix

I think your impression might be wrong

I was on a amiga-meeting recently with lots of hardware

Nothing "NG" except one mac mini with MorphOS in 30 minutes mode

and I was the only one using UAE

@OlafS25

Quote:

OlafS25 wrote: @amigadave one reason why X1000 was/is so expensive that R&D is covered by a small number of sales. It seems he has sold more than expected so I think he got back his expenses there. I doubt that this will be the case for X5000 but I do not know the calculation

That is incorrect. On more than one occasion, Trevor stated that the price of the X1000 did not include profits toward his initial investment in R&D for the Nemo motherboard. Wawa spreading rumors that it might have been a standard reference design board is ridiculous and would make me laugh, except that uneducated false statements like that can actually be harmful in some ways, so I don't think it is funny.

If you want to give opinions on what the X1000 cost, then you really should do more research into all statements made by all parties involved, and even actually research the cost of parts plus R&D to create such a new computer design. I think that making wild guesses about where the money has gone, or how much has been spent, without knowing first hand (or 2nd hand directly from statements made by Trevor, or Matthew Leaman), is irresponsible.

You are right that Trevor has sold more X1000's than he might have expected, but his share of the profits from that machine are quite small compared to it's over all selling price, so an extra 100, or even 200 X1000's sold are not going to make much difference in paying back the tens of thousands of dollars he spent on R&D for the X1000.

I am done with this topic of arguing with people who haven't done the research, or listened to what Trevor has been saying for the past several years. Believe what you want, but I know differently.

_________________
Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . .

@amigadave

im not spreading any rumours. i also have not said it was a reference design, but only might have been or might have been based on one. and i dont understand how that would be harmful. previous amigaone models were rebranded reference designs and it wouldnt harm them except that they were buggy.

@wawa

Quote:

im not spreading any rumours.

You're not spreading any facts either. A guess is the same as a rumor these days.

_________________
Hate tends to make you look stupid...

@OlafS25

Quote:

OlafS25 wrote: @itix But for what running 68k on PPC? The advantage of using PPC would be when using native PPC versions. And additionally many libraries are hardcoded asm partly direct hitting the hardware for speed reasons. That do not work on PPC except when using UAE (that has no JIT there).

Usually libraries do not directly hit the hardware. Do you some example?

Anyway, libraries can be replaced by native equivalents.

@OlafS25

Quote:

OlafS25 wrote: @itix I think your impression might be wrong I was on a amiga-meeting recently with lots of hardware Nothing "NG" except one mac mini with MorphOS in 30 minutes mode and I was the only one using UAE

Nevertheless, 68K hardware platforms are much limited, and very expensive if you add a fast accelerator board and/or some good amount of RAM.

New 68K platforms will come, for sure, but they are still expensive, limited, and have also very long time-to-market. Not even talking about specific drivers which are needed for taking advantage of (hopefully available) new/advanced features.

I personally see the "hardware way" too hard to satisfy customers need.

I think that a more modern and advanced Amithlon-like solution (replacing "sensible" APIs with native code, where possible) could much better fits the goals, providing also faster and easier development time.

@amigadave: that's only speculation, since there are no numbers.

But even taking the hypothetical ones that you reported, 200$ x 100 X1000 sold, means 20000$, which is close to the R&D costs that you talked about.

@itix

Quote:

itix wrote: @OlafS25 Quote: But for what running 68k on PPC? The advantage of using PPC would be when using native PPC versions. Wrong. The advantage is that OS is updated and developed. Due to historical reasons it is using PPC but it is just CPU arch. Native builds only have advantage in CPU intensive tasks and such tasks are quite rare. For example MP3 decoding is heavy task for 000-040 processors but for PPC 68k native MP3 decoding is not an issue. And of course the gfx and sound subsystem, datatypes, system libraries etc are native and optimized. This would not be case if MorphOS was 68k OS running in WinUAE.

That's the good part, but the bad one is that it also limits MorphOS and you cannot even go for a 64-bit PowerPC, otherwise you lose such cool feature.

Anyway, it's just a question of time: MorphOS will be ported to x64, so UAE is the only option to run old 68K apps (on a sandbox) without the same level of transparency & integration that it now offers now on PowerPCs machines.

Last edited by cdimauro on 04-Jun-2015 at 09:29 AM.

@amigadave

he is business man, I am propably not as successful as him but I studied economics. It is not magic but how the world works. That way price calculation works except you assume that he automatically calculated big losses when started the project. But if you know otherwise what shall I say. I can only offer human rationality and economics against it.

@cdimauro

FPGA and 68k of some sort (UAE, accellerator cards and so on) for amiga fans

A modernized platform for the rest. But you still need new software and new services for it that make it interesting for people because you are in competition to the big platforms then. A modernized platform without software will fail either. Don´t you think?

@sundown

Quote:

You're not spreading any facts either.

how much does that matter if x1000 is a reference design or not_ how much originality is there to be expected in the way cpu is connected to ram and pci bridge? aeon does not publish how much if any any return has been made or and how many units they have sold, so it is a non issue and may be treated as such. what matters to me are the results i can obvserve, not the company internals.

Quote:

A guess is the same as a rumor these days.

which days are that? people make assumptions all the time especially around here where there is so little open and dependable information. speculations are btw what happens most among os4 fans most of the time. and now you guys say its dangerous? simply because i made some observations? cmon.

@OlafS25

Quote:

OlafS25 wrote: @cdimauro FPGA and 68k of some sort (UAE, accellerator cards and so on) for amiga fans A modernized platform for the rest. But you still need new software and new services for it that make it interesting for people because you are in competition to the big platforms then. A modernized platform without software will fail either. Don´t you think?

Don't put things as mutually exclusive.

Amiga fans like 68K and use real Amiga (maybe accelerated expanded) and/or hardware simulators like Minimig et similia and/or PCs with (WinUAE) and/or PCs with PowerPC CPU (running EUAE and/or transparent emulation layers like Trance/Petunia).

But many Amiga fans use also AROS and/or MorphOS and/or OS4 with their native software.

Also, a modernized platform is required also for 68Ks, because that platform is limited too. That's if you like to have some evolution without being related to the 32-bit ghetto with its anachronistic 2GB ceiling.
Here AROS can help, because it can support a more powerful 68K clone, and even a 64-bit incarnation (there are ideas about extending the 68K ISA to 64-bit). But with 64 bits, you know, you are on the same boat of all other architectures: you must use UAE to run old 68K software in a transparent & integrated way (see Petunia and Trance).

Yes, new platforms need software. That's true. But you cannot think to stick to the old 68K binaries till the end of life. Otherwise, well, use your good old plain Amiga with its 880KB floppies attached to an old CRT monitor or TV.

@cdimauro

I already use a modernized OS on 68k

It is based on source code compatibility but if you would add things like memory protection that would not be the case anymore and that is what you talk about.

That will also the case for a possible MOS port to X64, it will mean not just recompiling but needing new software. I do not think that a new platform without software is a success. Who will write the software for it? On Aros you now have synergies between the different hardware platforms but that only stays as long the different platforms stay vastly API compatible.

@cdimauro

Quote:

But even taking the hypothetical ones that you reported, 200$ x 100 X1000 sold, means 20000$, which is close to the R&D costs that you talked about.

Trevor has said that R&D on the X1000 cost close to $200,000.

@OlafS25

Quote:

OlafS25 wrote: @cdimauro I already use a modernized OS on 68k It is based on source code compatibility

I think we use the same o.s./platform.
Quote:

but if you would add things like memory protection that would not be the case anymore and that is what you talk about. That will also the case for a possible MOS port to X64, it will mean not just recompiling but needing new software. I do not think that a new platform without software is a success. Who will write the software for it? On Aros you now have synergies between the different hardware platforms but that only stays as long the different platforms stay vastly API compatible.

Memory protection, SMP, resource tracking break compatibility even at the source code level, for sure, but that should happen at some time for every post-Amiga o.s..

In the meanwhile removing the 2GB limit is the first step which can be easily achieved. 2GB of TOTAL (o.s., devices, and applications) memory is really too low, because software increased complexity and requirements.

Especially for an Amiga-like o.s., were no MMU is usually used and memory is not allocated on a page basis, memory fragmentation increases and makes things even worse.

Anyway, AROS x64 is already available, and it's clearly the future mainstream platform such o.s., so more and more software will arrive (AROS, as you stated, is based on source-level compatibility, so what's already compiling for the 32-bit versions of the o.s. should compile also for x64, unless for badly written code). Keep focus on it.

@Rob

I do not think either that Trevor made lots of profit with X1000 even though selling more. I think most profit was made by Varisys. But I do not think that he was calculating 100.000$ loss when he started the project, he might be "crazy" but not as much as some people here seem to think . It is crazy anyway because he took a lot of financial risks with prefinancing it. He might be a enthusiast but he is still a business man. So I do not believe that at all. But we do not know that because we do not know costs and price calculations.

But it is his money anyway...

@Rob

Quote:

Rob wrote: @cdimauro Quote: But even taking the hypothetical ones that you reported, 200$ x 100 X1000 sold, means 20000$, which is close to the R&D costs that you talked about. Trevor has said that R&D on the X1000 cost close to $200,000.

This changes the things. amigadave talked about "thousands dollars". Here there's more than one order of magnitude of difference.

But still we don't know what's the real gain for Trevor.

Last edited by cdimauro on 04-Jun-2015 at 11:54 AM.

@cdimauro

for me staying source code compatible between the platforms is critical for a possible success, it is not only important for 68k but also general if you want to attract developers. It is also a benefit compared to Linux. Dropping it without having winning huge advantages would be a mistake in my view. But we will see.

Other than that we are not too far away

@wawa

Quote:

how much does that matter if x1000 is a reference design or not

It obviously matters a lot to you or you'd just accept what has been said by the people directly involved rather than come up with your own "possible" theory about some design sat in a drawer just waiting for the right customer to come along.