So what is XMOS good for?

[koaftder]

How come people already know the board will be underpowered? How come people already know the XMOS chip is useless?

I've been working with these chips for months. It's a neat little micro controller with cool architecture but it doesn't bring anything to the table on a desktop machine. There's nothing it can do that isn't better suited to doing on the host processor.



I'm not just trolling here, I'm actually using these things in projects.

[Vulture]

Isn't the XMOS chip this one:

http://en.wikipedia.org/wiki/XCore

according to wikipedia?

It seems pretty powerfull, at least on paper.
I, for one, would be happy if it could emulate classic miggy and keep this load (or a big part of it anyway) away from the CPU. I've no idea if it'd be possible, but that would be a useful utilization.

Also, afaik, PPC chips are still developed and manufactured under the PowerISA specs which is updated fairly regulary. Price-wise, well, I don't expect it to be cheap, but the 2000$ mark seems too high. Of course I can see the train of logic leading to that figure, but I think it's gonna be much cheaper than that. We'll see....

[Matt_H]

What's all this negativity about? Puzzles me....
How come people know the price already before it's announced? How come people already know the board will be underpowered? How come people already know the XMOS chip is useless? I mean, all these could indeed be true, but - well - they also could be false assumptions. So, back to my original point, why all this negativity? *scratches head*

In a way, they're sort of inviting this upon themselves by burning the momentum they generated with the puzzles back in January and dancing around the actual complete specs of the machine.

Still, I agree. This thread is just a liiiitle too hostile. Let's see what can be done with it in a real-world implementation. Or, if it must be put another way, give them enough rope to hang themselves and see what happens.

[dammy]

I'm thinking of porting the compiler over to AROS, just to be a dick, just so AROS users can brag that they got to play with the cheap XMOS dev kits first without paying 2,000 for the privilege of having the uber l33t new hyperion OS/mobo combo and still have to compile their junk on a mac or pc.

https://www.xmos.com/technology/design-tools-source

I thought it was all closed sourced tools, nice to see I'm wrong.  How big of a job would it be to do the port?

[unusedunused]

>Its definitely good for looking busy, and it gives you the ability to say your computer >uses "custom chips" when

The xmos is no custom chip.Its a mass ware that is not only available in the X1000

A custom Chip is more common called as a ASIC

http://en.wikipedia.org/wiki/ASIC

The sense of the xmos in the X1000 mainboard i cant understand.I read in the docu and as far i can see when use maximum of 5 wires the Chip can transfer max 800 megabit /Sec.
(page 9/26)
http://www.google.de/search?q=xmos+link+speed&rlz=1I7RNTN_de

So this make in best case a Interconnect rate of 100 megabyte/sec.Thats little slower as a PCI and USB3 is fast enough to connect the Chips.

But the transfer speed of the xmos is way too few to do some acceleration, because even a slow PPC CPU as the SAM can load and store data at 250 Megabyte /Sec

also when somebody want use the XMOS, he need a controller card that offer the IO.Maybe this should put in the Xorro Slot.

But when this should put in the Xorro slot the xmos Xhip can also put on the Card.this make the X1000 little cheaper and need little less power.

[Piru]

I, for one, would be happy if it could emulate classic miggy and keep this load (or a big part of it anyway) away from the CPU.

It can't.

[Tripitaka]

Newsflash POWER PC is dead. You can add all the off the shelf components you want to it, but its still dead, dead dead at its core...
 
Steven

Looked inside a PS3 did you?  .....hmmm, well, what did you see? Oh yes, one of them. Don't look dead to me buddy! Got a 360? yeah, in that one too......geesh, glad your not my doctor.

[Tripitaka]

@ Piru

Please justify your comments. "No" is meaningless without an explanation.

[Tripitaka]

@ Koaftder

Cool stuff, good to see someone on the thread with some hands on knowledge of the chip.
So, rather than dwelling on what can't be done, please tell us a bit about your experience,  what have you done with the chip so far?

[Piru]

@ Piru

Please justify your comments. "No" is meaningless without an explanation.

Justify? You mean "elaborate" I guess.

Right here we go:

1. The device has 64Kb of RAM. That is way way too little for most tasks. Certainly too little for any kind of emulation.
2. The device has no direct access to system RAM (the access is really slow and software aided. The device doesn't cache such memory accesses). Emulation needs to have fast access to the state information.
3. Emulation in general isn't suitable for threading. That's why we don't (and won't) have any dual core support in WinUAE, for instance. Thus, having 8 or more cores doesn't make any difference.
4. Trying to use XCore for some parts of the emulation would probably only end up slowing the whole down. It'd be much faster to do the whole thing with the main CPU (which has direct fast access to system memory, with L1 and L2 caches).

These are the obvious reasons why XMOS sucks at emulation. Same reasons make many other things unsuitable, aswell. For example using the device for some kind of graphics or audio acceleration.

I haven't even looked into instruction set and such issues, yet. What is clear however that there is no FPU, which means any floating point math would need to be done with integers (possible but very slow).

[Tripitaka]

Thanks Piru, that's a fair answer. Looks like an NVidia or ATI chip may have been a better choice, at least CUDA, physX etc.. have some worthwhile uses.

[SamuraiCrow]

I'm thinking of porting the compiler over to AROS, just to be a dick, just so AROS users can brag that they got to play with the cheap XMOS dev kits first without paying 2,000 for the privilege of having the uber l33t new hyperion OS/mobo combo and still have to compile their junk on a mac or pc.

https://www.xmos.com/technology/design-tools-source

It's an LLVM-based toolchain!  I hope you do port it because I want to make Mattathias BASIC use that same toolchain!  The trickiest part should be that their IDE is Eclipse and requires Java.  Other than that, it's just a matter of making the calling conventions for AROS libraries using TableGen and setting the alignment for AROS' structure alignment policies.  There's already an x86 and PowerPC backend for LLVM but x86 seems to get more attention.

[haywirepc]

Tripitaka
 
Okay, except for embedded devices and video game consoles... which use 2+year old slower chips...Power pc is dead dead dead.
 
From wikipedia :
 
"In 2004, Motorola exited the chip manufacturing business by spinning off its semiconductor business as an independent company called Freescale Semiconductor. Around the same time, IBM exited the embedded processor market by selling its line of PowerPC products to Applied Micro Circuits Corporation (AMCC) "
 
What do you think about the future of power pc without the millions and millions of R+D by IBM and Motorola GONE? Intel/AMD wins hands down and thats that my friend. Freescale and applied micro circuits corporation don't have 5% of the budget of motorola or ibm.
 
They can not and will never again compete with intel/amd, no way.
 
I for one, can't understand anyone gambling the future of their company
on a dying processor. Its dead or dying. It will NEVER be the competitor it once was, NEVER. You can be ignorant of this and put your blinders on
but that fact will NEVER change.
 
I for one would buy AOS in a hearbeat if it ran on x86 and so would so many others, yet they still ignorantly think thats not a good idea.
 
Also, I love "Our most ambitious project to date" If the most ambition they can muster is attaching an off the shelf microcontroller to a substandard overpriced power pc motherboard... I don't know thats kind of sad really.
 
Somehow I hoped for more ambition than that...
 
Steven

[yssing]

Haywirepc >> I think I remember a thread about not doing drugs..

PowerPC is so far from dead. Its true that the consoles mentioned, PS3, Xbox, Wii are a few years old, but they all uses PowerPC¨, they are still being made, and you get a lot of power, very cheap.

IBM uses PowerPC in some of their servers.

But I guess, that its easy to claim that PPC is dead, when your agenda is to have AOS4.1 on X86. Go buy the iMica, its cheap, its AOS, sort of, and its X86..

@all I can definetly see some very good uses of the xmos. Like integrated robotics interface, advanced gate controller plus lots more. I am sure its some thing the company I hope to start can use.

I am excited about this.. honestly..

I am very tired of reading all thsi negative stuff over and over againg.
A saying gose some thing like: "if you have nothing nice to say, dont say anything"

[dammy]

@all I can definetly see some very good uses of the xmos. Like integrated robotics interface, advanced gate controller plus lots more. I am sure its some thing the company I hope to start can use.

What does integrated robotics or advance gate controller have to do with end users and their desktops?  Or for that matter on a pure hobby level, why wouldn't USB be enough for a hobby robotic controller?

[mongo]

Tripitaka
 
Okay, except for embedded devices and video game consoles... which use 2+year old slower chips...Power pc is dead dead dead.
 
From wikipedia :
 
"In 2004, Motorola exited the chip manufacturing business by spinning off its semiconductor business as an independent company called Freescale Semiconductor. Around the same time, IBM exited the embedded processor market by selling its line of PowerPC products to Applied Micro Circuits Corporation (AMCC) "
 
What do you think about the future of power pc without the millions and millions of R+D by IBM and Motorola GONE? Intel/AMD wins hands down and thats that my friend. Freescale and applied micro circuits corporation don't have 5% of the budget of motorola or ibm.
 

IBM exited the embedded processor market not the PowerPC market (IBM has just recently announced the new PowerPC A2), and while I don't have the numbers, I would imagine Freescale spends about as much on semiconductor R+D as Motorola did.