Xmos interpretation issue

[DanB]

btw Folknology,
where did you got this code from ? (could you post full code?)

I think Folknology is just raising, in a humerous way, that an XS1 should not be considered a microcontroller (MCU)... May be wrong though! :)

[jonathan]

I think I might have found your problem Al. Read this article:

http://www.electronicsweekly.com/Articl ... alyser.htm

Electronics Weekly ignores the term microcontroller from the press release, favouring instead microprocessor/processor. Clearly, they don't know what they are talking about, but needless to say, even if you did mean to test equality, this seems to point to the fact your conditional can never become true.

Hope that clears this up. Perhaps XMOS can close the ticket you sent in now.

[octal]

DanB killed this topic
You'll be banned forever if you continue :ugeek:
:mrgreen:

[DanB]

DanB killed this topic
You'll be banned forever if you continue :ugeek:
:mrgreen:

Do'h. I genuinely thought you did not understand... And instead the last laugh is on me! :oops:

[Folknology]

Jonathan : Clearly Electronics Weekly are still not getting the message, Xmos may have to send the boys around to persuade them!

Octal: Event Driven Processor - There you go, they can ban me now, I will probably keep using it ;-)

DanB: Your not wrong...

Code: Select all

Assert (XS1==MCU) -> false // clearly not true
Assert (XS1=MCU) -> false // type coercion problem
Includes(XS1.features,MCU.operation) -> Maybe true // possible intersect of features depending on operation

Just a bit of fun folks

regards
Al

[Berni]

Well a xmos chip can still be used as a very fast microcontroler, altho the large amount if support circuitry does turn you away some .Usually you just give a MCU 3.3V and it starts running (This is why i love them) But a xmos can do far more than a MCU.

[octal]

hi Berni,
(THIS IS MY OWN OPINION)
I personally think that "if you want to replace an MCU with an XMOS, you are simply wrong" !!!
In most cases, an XMOS is a better replacement for an FPGA, or an FPGA+MCU than for an MCU.

XMOS are not simple microcontrollers. They have microcontrollers functionnalities, and missing functionnalities (at least digital ones) can be created with software. They are fast, extremely fast! I agree.
But XMOS are special kind of deterministic parallel machines, they are simply "another" category of tools. They are not there to replace other microcontrollers, they are not there to replace FPGA, they are there to fill the gap between these two besties! having the speed, parallelism and determinism of FPGA, the prorgammability of an MCU and the flexibility of microcontrolled logic!

As for complexity of supporting circuitery, I would say that it's not that complex, it's simply that XMOS chips are a professionnal grade chips that need some expertise and a lot of prototyping tools mainly available at professionnal offices. Hobbists will for sure not find them easy to manipulate.

As for usability, this is another problem. Users mainly tend to use and abuse of using the same microcontroller (at least in the same familly) in ALL their designs. Users experienced with PIC chips, will want to use them for everything. Idem for AVR or ARM chips. This is a natural habbit for humans, and economically this is ONE OF the best ways to become profitable: you reuse your libs, codes and experience.
With technology evolving quickly these last years, and product becoming obsolete in less than 3 or 4 months, doing things this way I think is wrong! Users (developpers) should forgot their sentimental links to their beloved tools, and choose to use the best and correct tool for the developed solution.
If I had to make a text console on a little alphanum LCD for example, connected through USB or RS232, I would certainly not use an XMOS (unless I have spare pins and mips and the console need to be embedded in the desing). I would use a cortex-M0 or such low cost embedded MCU (less than $1 compared to the $5 XMOS).
If I had to make a camera controller or a graphic LCD controller, instead of using an FPGA I would use an XMOS, as for such solutions we need to control SIGNALS TIMING precisely and we need the processing capabilities of an MCU, and we need SPEEEEEEED. Instead of cascading an MCU with an FPGA, an XMOS is a wonderful solution for such problems. In this case, adding 1 or 2 more components to handle power management stuff is really not a ptoblem at all.

I think that most users make the assumption XMOS==MCU (and never do FPGA==MCU) mainly because XMOS are programmable using almost a language (XC) that is close to "normal" languages used for MCU.
XMOS made the great thing of making a PROGRAMMABLE reconfigurable silicon (like an FPGA) having the processing capabilities of an MCU and prorammble using almost the same tools despite the fact that it's a parallel architectured machine.

I think that XMOS answers a real market need, in the fact that most of the time when we need complex logic, we use FPGA+MCU. Most of the time FPGA alone is not useful because we need some complex filters and some complex logic to process data. If data are to be processed at the speed of an FPGA, we need to cascade it with a quick (power hungry) MCU. When processing capabilities needed are not that important, some FPGA makers provide some SOFT processors cores implementation (ZPU is an example). But when you need extreme flexibility, and when data should be trasfered multiple times between the "intelligent part" of the board (the MCU) and the extremely fast hardware logic (FPGA), you have to deal with all data transfert complexity and effectiveness. XMOS solved all that by providing an integrated solution: a software reconfigurable programmable logic that can manage all FPGA stuff using SOFTWARE reconfigurable ports, and a programmable intelligent core (like an MCU), all that with the goodie of being parallel and implementing some extremely quick data channels (XLINKS) : all either in the same chip or with pluggable separate chips.

[jonathan]

Great post octal.

[russf]

It's not April Fools' Day, but Al was having some fun with us.

We should not take anything he says seriously, in future.

--r.