@Hyperionmp

Quote:

The integration is pretty tight.

Tease We want more info!

Quote:

Quite frankly, I fail to see why putting the 7 USD XMOS chip onboard as opposed to inserting it in USB or PCI slot is a big deal.

Oh, I agree with this and everything else you wrote. The additional cost of it is too low to deter any dedicated Amiga users, and any new Amiga needs to have something "different" to grow the user base if there's going to be any hope of growing the community again...

It's pretty clear the Amiga won't compete on raw performance again for a very long time, if ever, and so it needs these little things to attract users for other reasons than price/performance.

I mean, look at Apple - the price/performance of a Mac is ridiculous compared to a Windows or Linux box. I recently bought two laptops for my wife and myself for about 30% less total (for both of them combined) than it would cost to replace the single Macbook Pro that (had to have it in my last job) it replaced. The news ones are faster, with more memory and larger disk... If anything, Apple is proof that it's possible to survive and grow with ridiculous markups in a markups if people feel the *experience* is better.

That's where a new Amiga has the best hope of adding new users: Provide an interesting and unique experience.

The XMOS is a nice experiment in that regard. Maybe it won't succeed, but we won't know if we don't try :)

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

@Hyperionmp

Hi and thank you for your message, but could you please elaborate more on "Integration is pretty tight"?

Does it mean the XMOS chip(s) is integrated to the motherboard as the main CPU is (it has access to any part of the system)?

Does it mean the OS is fully aware of it and it's software support can be done OS-Friendly?

Or maybe both?

_________________
God created Paula so that Allister Brimble and Dave Whittaker could do music

Check my Amiga gameplays (ITA)!

@Hyperionmp

Quote:

Hyperionmp wrote: If that logic were followed, all I/O should be located on a PCI, PCI-E or USB bus because you have SATA, USB 2.0, ethernet, soundcards etc. based cards/controllers which fit into these slots. Why is anyone still making Southbridges?

SATA or usb is used by everyone each day, but the XMOS chip hardly belongs to those things used every day. It is a uC for special purpose, mostly time critical I/O.
A mainboard that will be successful has to have all *everyday things* onboard and offer the option to customize it by adding hardware yourself. But you have to keep the trade off between adding features (and increasing production cost) and covering as many functions as possible. That is one of the first lessons you learn in economics.
A feature that offers no benefit for the average user but increases cost, is contra productive. You may try to market it as an outstanding feauture, but then you have to prove what is really outstanding about that feature (development tools? reliability of the host system? integration to existing solutions? Migration and support offerings?). And still I fail to see the outstanding. I.e. I doubt that the addition of the XMOS chip will drive sales more than it actually avoids due to an overall expensive dsign (KISS!). But it is your endeavour and if you really come around with an outstanding feature I will be more than willing to milk my bank account to buy a board or two. But yet I do not see the demand on my side.

_________________
My programs: via.bckrs.de
MorphOS user since V0.4 (2001)

@Thread

I just downloaded the dev tools to play with from the Xmos site, thier dev environment
is called 'Workbench'


Wol.

Last edited by Wol on 14-Jan-2010 at 12:12 PM.

_________________
It is my conviction that killing under the cloak of war is nothing but an act of murder.~Albert Einstein

@Hyperionmp

Quote:

Quite frankly, I fail to see why putting the 7 USD XMOS chip onboard as opposed to inserting it in USB or PCI slot is a big deal.

To be perfectly honest I was a bit disappointed that the single core chip was chosen and not the quad core, considering the small cost.

_________________

@Zardoz

Yeah, but what if its main purpose is to act as a bridge to the X1000's main memory for their 64 Core boards. It would effectively contain all the same circuitry to facilitate this. Anyone know of any XMOS development boards that can access external RAM, as this could give us a clue as to how they seek to accomplish this on the X1000?

_________________
Leo Nigro, CTO Commodore USA, LLC
Opinions expressed are my own and not those of C= USA.
Commodore/AMIGA "Beautiful, High-Performance, Home Computers for Creativity and Entertainment."

@Zardoz

Quote:

Zardoz wrote: @Hyperionmp Quote: Quite frankly, I fail to see why putting the 7 USD XMOS chip onboard as opposed to inserting it in USB or PCI slot is a big deal. To be perfectly honest I was a bit disappointed that the single core chip was chosen and not the quad core, considering the small cost.

Couldn't agree more. Why the 7 USD (that's less than 5Euros ONLY) was chosen when you could add 15 euros more and have the best available xmos chip..

_________________

Double realtime:

AOS is near realtime (desktop) OS (unlike any other).
With XCore Amiga X1000 will be capable to handle Hard Realtime tasks withoput any affect on the desktop OS.

To me it sounds like double realtime machine.

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

@BigBentheAussie

They don't have a memory bus, so the only alternative (at least with the chips currently listed for sale) is to implement a RAM controller in software running on one of the threads. There are example memory controllers available on their site, but that approach pretty much makes it impossible to get high transfer rates to/from main memory. It's not the kind of use the chips are designed for.

I agree with Zardoz - the 4 core version would've been very welcome. But I guess a cheap extension card with more cores won't be hard to do if/when we have some apps that make use of the single onboard core.

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

@KimmoK

An operating system is either real-time or it isn't. There's no "near real-time". To qualify as real-time, interrupt events and the like (whether asynchronous or from timer) must be dealt with within a specified time limit. Any operating system that makes no such guarantee is not real-time.

AmigaOS, under ideal conditions, is fast and responsive, but it sure isn't real-time.

Last edited by Karlos on 14-Jan-2010 at 02:21 PM.

_________________
Doing stupid things for fun...

@vidarh

From the XMOS site:
http://www.xmos.com/applications/memory/sram-controller

1 thread of the G1 at 50 Mghz provides up to 30 MB/sec
Can it use all 8 threads?
Is this fast or slow?

Just for comparison, what would it be on an typical x86?

Edit: Ah. I see it should have been in the order of 3000+ MB/Sec for RAM.
Edit: The speed is actually more comparable to reading from an SD Card.
Edit: Am I looking at the wrong thing?

Last edited by BigBentheAussie on 14-Jan-2010 at 02:46 PM.

_________________
Leo Nigro, CTO Commodore USA, LLC
Opinions expressed are my own and not those of C= USA.
Commodore/AMIGA "Beautiful, High-Performance, Home Computers for Creativity and Entertainment."

Btw. to not just go back and forth with the same arguments endlessly, here's something more constructive:

Here's a pretty cool blog written by a guy that's planning on hooking up audio, SD card interface, IDE and memory to his XC2 development card to build an audio player. Fun little project to read about. It's not directly applicable, since it's a standalone dev board, but his various solutions give some indication of how easy it is to interface various things to it.

And here's someone doing a low latency audio project, claiming he gets 70usec latency and that the likely BOM for an XMOS based card to provide it will be on the order of $15 compared to $3000 for the solution he currently owns.

If they have the funds for it, A-Eon ought to offer a few guys like these steeply discounted or free X1000's in return for coming up with some interesting projects...

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

@KimmoK

Quote:

AOS is near realtime (desktop) OS (unlike any other).

You just don't know what RTOS is.

_________________
OpenWindows Initiative. Port PS3 hardware to bananas. For free. Join today and receive expired $50 cupon from AI!

@Karlos

Quote:

An operating system is either real-time or it isn't. There's no "near real-time". To qualify as real-time, interrupt events and the like (whether asynchronous or from timer) must be dealt with within a specified time limit. Any operating system that makes no such guarantee is not real-time.

To nitpick a little bit, a system that "almost always" meets guarantees is considered soft realtime, so it's not strictly true that events need to meet a specified time limit all the time in order to be realtime. OTOH to be considered "soft realtime" a system generally have to provide gurantees about how it will handle failure to meet response times (i.e. it might guarantee that it will drop them, and that it will never drop more than, say, 5% of the interrupts).

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

@thread

You know what would be awesome?




Connect Xena with the A1200 clockport and get in control of the old lady over the new girlfriend.



I know that might not work, but I like the idea....

_________________
Bite Me!

@brotsalami

Quote:

Connect Xena with the A1200 clockport and get in control of the old lady over the new girlfriend.

Now you're just being dirty

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

@Karlos

"AmigaOS, under ideal conditions, is fast and responsive, but it sure isn't real-time."

You just described what is "near real-time".


(real-time as a definition is very flexible, btw. It becomes very clear if one goes through some tens of RTOSs evaluation)

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

@KimmoK

Now try AmigaOS under heavy CPU load. Not quite fast and responsive any more, is it?

_________________
Doing stupid things for fun...

@opi

But I think I know.

(but it does not mean that I might be over simplifying things)

One thing is sure. 68K AOS has been the only desktop OS that can get pretty close to soft realtime specs/needs. I think that is one of the core reasons of why NASA (and many others) have used Amigas.



btw from wiki: A real-time operating system (RTOS) is a operating system (OS) intended for real-time applications. Such operating systems serve application requests nearly real-time.
...
A real-time OS that can usually or generally meet a deadline is a soft real-time OS, but if can meet a deadline deterministically it is a hard real-time OS.

http://en.wikipedia.org/wiki/RTOS
btw. this RTOS was modelled after AOS: http://en.wikipedia.org/wiki/Nokia_OS (I worked with one of the core developer)

But perhaps talking AOS as a soft RTOS would be better and more appropriate than any "new" term. Anyway. X1000 handle both. Easily. As a standard. (hard realtime might need some downloads from XMos)

Last edited by KimmoK on 14-Jan-2010 at 04:13 PM.
Last edited by KimmoK on 14-Jan-2010 at 04:10 PM.
Last edited by KimmoK on 14-Jan-2010 at 04:09 PM.

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

@Karlos

CPU load does not affect AOS at all.
(except if priorities are not set properly, same with every RTOS)

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