Go Down

Topic: Arduino Due with SAM3U4E (Read 13015 times) previous topic - next topic

fat16lib

Wow what a change. 

http://www.i-programmer.info/news/91-hardware/3081-arduino-goes-arm-new-modules.html

What features should the core system software have?

Hope it is a true RTOS with tasks.

It would be great if the base was a proven open source RTOS from this list http://en.wikipedia.org/wiki/List_of_real-time_operating_systems

I have use several of these.  I like ChibiOS/RT, it runs on ARM Cortex M3.

retrolefty

Quote
Hope it is a true RTOS with tasks.


My bet would be that there be no standard RTOS as part of the Arduino IDE support of this new board. But just rather the same level of support the the existing boards have, compiler/linker/libraries, etc. However 3rd party will certainly address adding such a OS, as the boards resources can certainly handle that.

Lefty



fat16lib

retrolefty,

Your right, the Arduino team could never produce the kind of software required for this class of processor.

This is a "Live Free or Die" moment.  It will be easy to have great software independent of the Arduino team.

I would start with the Eclipse IDE and YAGARTO.

YAGARTO is a cross development environment for the ARM architecture, running on a Windows host. It includes the GNU C/C++ toolchain and the Eclipse IDE.

I would then us one or more RTOSs that I have ported before.  ChibiOS/RT, FreeRTOS, NuttX RTOS, and others.

All we need from Arduino is cheap mass produced hobby style hardware.


pjrc

But will it be "cheap"?  Has any pricing been announced?  What if it ends up being in the same price range as Arduino Mega?

fat16lib

Since Atmel is doing it with Arduino I have hope for cheap.  Atmel is talking about it in news releases.

What price would be right for a board with this part?

westfw

Quote
What price would be right for a board with this part?

The chip is slightly cheaper than the ATmega2560, and I believe the idea is that the separate USB/Serial chip will also be eliminated.  So I'm hoping for "slightly cheaper than the current MEGA2560 board."

Quote
It will be easy to have great software independent of the Arduino team.

No, it won't.

Quote
the Arduino team could never produce the kind of software required for this class of processor.

If it were easy, the Arduino team would be able to do it.  They mentioned (in the OSHW summit speech) not needing venture funding because they were doing pretty well at being self-funded.  Well, one excuse for venture funding is the need to suddenly expand the size of your company by hiring a bunch of people...

Quote
All we need from Arduino is cheap mass produced hobby style hardware.

So where is the Arduino Abstraction layer for the OTHER cheap, mass-produced, ARM hardware?  MBed, STM32-explorer, LPCXplained, Kinetis, Stellaris boards, etc?  It shouldn't require a shield-compatible board to have started with all that software work, and you can argue that the Arduino hardware team isn't the ideal "mass production" facility just the way you can claim that the Arduino SW team isn't up to a major SW transition...

I am looking forward to having a reference implementation for how an on-chip USB perihperal would be used within the Arduino environment (requiring some sort of cooperative tasking beyond what currently exists, if not an actual RTOS.)  But I don't think that it will be "easy", or even close to easy...

fat16lib

Westfw,

I have no interest in an Arduino layer.  I like existing RTOS systems for ARM.  They already exist and I use them.  That's easy.

It takes a lot of talent and effort to build a good RTOS.  The ideas for a 8-bit avr won't cut it.

I own alot of the OTHER ARM hardware.  They are evaluation boards or stuff made by small companies.

Arduino seem like a good new option for ARM hardware for hobby use.

bperrybap


I am looking forward to having a reference implementation for how an on-chip USB perihperal would be used within the Arduino environment (requiring some sort of cooperative tasking beyond what currently exists, if not an actual RTOS.)  But I don't think that it will be "easy", or even close to easy...


But we kind of already have that today. Go have a look a Paul's teensyduino USB implementation.

--- bill

anon10500



I am looking forward to having a reference implementation for how an on-chip USB perihperal would be used within the Arduino environment (requiring some sort of cooperative tasking beyond what currently exists, if not an actual RTOS.)  But I don't think that it will be "easy", or even close to easy...


But we kind of already have that today. Go have a look a Paul's teensyduino USB implementation.

--- bill


Isn't it closed source ?

bperrybap

Teensyduino is open source. So is Paul's non arduino C code USB device code.
It is only Paul's halfkay bootloader that is closed source.

--- bill

frank26080115


Teensyduino is open source. So is Paul's non arduino C code USB device code.
It is only Paul's halfkay bootloader that is closed source.

--- bill


The PCB layout is not open source either.

tochinet

Two months later, many people interested in Due, is this a stillborn thing ? Or good news for Xmas ?

graynomad

It has gone very quite on the Due front, I gather there are beta testers working on them though. I just searched the 1.0 code and could find no reference to it in the pins file.

I wouldn't make room in your xmas stocking.

______
Rob
Rob Gray aka the GRAYnomad www.robgray.com

pjrc

On the developers mail list, Massimo Banzi of the Arduino Team recently wrote:

Quote

the Due is in private alpha at the moment.
There will be a public beta in january


Unless you can somehow get into their private alpha program, looks like Santa will be bringing your Due a few weeks later.

There doesn't seem to be any mention of Due at all in Arduino's github.  But then, when Leonardo was announced, suddenly lots of files became visible, complete with edit history in github stretching back at least several weeks.  So it's possible a lot is actually going on, but none of us can see it?

There was mention, a few months ago, about an experimental IDE based on Qt... but nothing mentioned since.  It's entirely possible Due might have a completely different IDE?


Regarding "a reference implementation for how an on-chip USB perihperal would be used within the Arduino environment", indeed Teensyduino (I am the author) has provided such code for nearly 3 years, with open source MIT licensing.  Indeed, the PCB layout and bootloader are not open source... so if you want a development board which you can clone with zero effort and resell in competition with the original project (as many unscrupulous Asian companies do), then Teensyduino not a good fit for you.  But if you want an open source "reference implementation for how an on-chip USB peripheral would be used within the Arduino environment", all that code used within the Arduino environment freely available under the extremely permissive MIT license, and has been for almost 3 years!

Of course, the Leonardo code is now part of Arduino 1.0.  It uses a much prettier C++ syntax, which you may prefer?  Or maybe you'll like my messy but "everything right there" C-only style?  Leonardo's code shares (perhaps borrows?) a lot of Teensyduino's design, as opposed to the way things are done in Atmel's reference and Dean Camera's LUFA library.

If you want open source code to see how this works, there's plenty of it out there for you.

westfw

Quote
"a reference implementation for how an on-chip USB perihperal would be used within the Arduino environment", indeed Teensyduino [provides that]

Yes, but we seem to be in a state where having a design be officially "blessed" and adopted by the Arduino team is going to be more important (market-wise) than "prior art" or technical capability.  And the Arduino team has their share of NIH syndrome, and/or issues to address that are not readily apparent to "the larger community."  What the Arduino team does OFFICIALLY WRT multiple CPUs and on-chip USB is going to be more significant (dollar-wise?) than Sanguino or Tiny or Teensy or ChipKit...

It makes the politics behind open source software (where real money need not be involved) look like a kid's lemonade stand.

Go Up