Author Topic: Intel Edison  (Read 17607 times)

0 Members and 1 Guest are viewing this topic.

Offline reasonTopic starter

  • Newbie
  • Posts: 6
  • Country: bg
Intel Edison
« on: September 23, 2014, 07:44:53 am »
Hi everyone, just saw the Intel try to enter the hobby microcontrollers  market with this product https://communities.intel.com/community/makers/edison/getting-started/overview, its pricey but comes packed with integrated wi-fi and bluetooth.
It will be interesting to see a review of this product from Dave  :D
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #1 on: September 23, 2014, 08:28:01 am »
More info here:
https://www.eevblog.com/forum/microcontrollers/intel-edison/

But to me, it seems they (Intel) don't have their stuff together just yet.

 

Offline reasonTopic starter

  • Newbie
  • Posts: 6
  • Country: bg
Re: Intel Edison
« Reply #2 on: September 23, 2014, 08:48:34 am »
didnt know the topic already exists, its already selling, breakout boards and the MCU board, this thing comes with a lot of memory , the price is kinda ok considering the extras that comes with. the neat things it has an arduino compatible breakout board
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #3 on: September 23, 2014, 08:53:59 am »
The problem is the GPIO speed not the headers, the Galileo is even worse at 230Hz on most GPIOs

My Cypress pioneer dev kit at $25 has Arduino headers and can use them way better than the Edison or Galileo

Although I don't have an Edison and it seems more competent than the Galileo.
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Intel Edison
« Reply #4 on: September 23, 2014, 11:37:17 am »
What is it designed for?

Sounds interesting but I stopped reading at the Arduino IDE part. With a real IDE, it may be worth exploring.
================================
https://dannyelectronics.wordpress.com/
 

Offline Rasz

  • Super Contributor
  • ***
  • Posts: 2616
  • Country: 00
    • My random blog.
Re: Intel Edison
« Reply #5 on: September 23, 2014, 08:34:00 pm »
What is it designed for?

to mark a KPI checkbox on some Intel managers spreadsheet
Who logs in to gdm? Not I, said the duck.
My fireplace is on fire, but in all the wrong places.
 

Offline westfw

  • Super Contributor
  • ***
  • Posts: 4199
  • Country: us
Re: Intel Edison
« Reply #6 on: September 23, 2014, 11:40:09 pm »
Quote
What is it designed for?
"Internet of Things", of course.  At least it has a credible amount of hardware and a real internet-capable connection, unlike many of the "It's small and has a communication interface!  Ideal for IoT!!" marketing of some vendors.  I've never quite understood all the people spending $30 for an Arduino (with 32k of program memory) and another $70 or so for a WiFi Shield of some kind.
 

Offline reasonTopic starter

  • Newbie
  • Posts: 6
  • Country: bg
Re: Intel Edison
« Reply #7 on: September 24, 2014, 06:50:56 am »
What is it designed for?

Sounds interesting but I stopped reading at the Arduino IDE part. With a real IDE, it may be worth exploring.
Arduino IDE is more than enough for most hobby work , if you need more professional stuff there are other solutions, but i think its only device in this size that comes with 1gb ram , 4gb for storage, wi-fi and bluetooth, on it you can run everything like home brew linux, android, custom firmware, whatever. I'm glad that the times for spending $70 or more just for a wi-fi shield are about to end. for many it could be the main choice because of this.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6460
  • Country: nl
Re: Intel Edison
« Reply #8 on: September 24, 2014, 07:17:58 am »
its pricey 
Quote
What is it designed for?
"Internet of Things", of course.
Two very opposite quotes there.
It looks like Intel wants a piece of the raspberry pie. If they don't market it below the pies price they don't get it and won't get it.
 

Offline PointyOintment

  • Frequent Contributor
  • **
  • Posts: 327
  • Country: ca
  • ↑ I scanned my face
Re: Intel Edison
« Reply #9 on: September 27, 2014, 12:01:37 pm »
What is it designed for?

Sounds interesting but I stopped reading at the Arduino IDE part. With a real IDE, it may be worth exploring.
It runs full Linux; you can write whatever Linux app you want.

Quote
What is it designed for?
"Internet of Things", of course.
Two very opposite quotes there.
It looks like Intel wants a piece of the raspberry pie. If they don't market it below the pies price they don't get it and won't get it.
SparkFun's rebuttal
I refuse to use AD's LTspice or any other "free" software whose license agreement prohibits benchmarking it (which implies it's really bad) or publicly disclosing the existence of the agreement. Fortunately, I haven't agreed to that one, and those terms are public already.
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6716
  • Country: nl
Re: Intel Edison
« Reply #10 on: September 27, 2014, 07:54:23 pm »
The problem is the GPIO speed not the headers, the Galileo is even worse at 230Hz on most GPIOs
Galileo v2 has 12 native GPIOs right? Are they still slow?
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #11 on: September 27, 2014, 08:13:16 pm »
The problem is the GPIO speed not the headers, the Galileo is even worse at 230Hz on most GPIOs
Galileo v2 has 12 native GPIOs right? Are they still slow?

Not sure, I only have a Rev 1 board :(
But I bet they are faster than the cypress i2c solution they used in the Rev 1 board.

If truly native and the board runs at 400MHz (and I found mine running a bit over 405MHz) then it should be able to do up to 200MHz on the pins if they are all on the same address.

Here is a link that includes the code I used to read the time stamp counter:
https://www.eevblog.com/forum/microcontrollers/intel-quark-soc-running-windows-iot-on-intel-galileo-(rev-1)/

Of course Microsoft doesn't yet support the Gen 2 boards so I'm not going to get one because I really don't want to use the arduino IDE.

The Edison IO is restricted by the Quark processor running at 100MHz so the Edison module can drive the I/O at 50MHz tops.

No news yet if the Edison will be supported by Microsoft IoT windows.
 

Offline johnnyfp

  • Frequent Contributor
  • **
  • Posts: 261
  • Country: nz
Re: Intel Edison
« Reply #12 on: October 08, 2014, 09:13:50 pm »
Quote
The Edison IO is restricted by the Quark processor running at 100MHz
What makes you say that? I can't see anywhere that says the Edison is restricted by the Quark MCU for IO. As far as I can tell the Quark MCU is yet not in use or available for use until Intel provide some drivers for it for the Yocto Embedded linux to use.

Plug time. I've made up one of these Edison Simple Breakout board so have direct connection to the IO and can test and see how fast I can get the IO to actually output, but in all honesty do you really need more that 50Mhz switching speed on a GPIO?
Would not the SPI or SDIO be better suited to doing anything at a higher speed?

 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #13 on: October 08, 2014, 09:48:19 pm »
I still can't find a diagram of the actual Edison and how things are connected.
So the Quark is disabled right now?

Found some more info here, earlier I did hear that the Quark handled the I/O, but since there is no diagram I can't tell either way.

http://linuxgizmos.com/edison-iot-module-ships-with-atom-plus-quark-combo-soc/

Quote
we have a single module with a single Tangier SoC that combines a modified, dual-core Merrifield Atom running Yocto Linux with a single-core Quark that is initially disabled. On the Galileo, the Quark runs the show, but on the Edison, it appears to be relegated to a microcontroller-like role, with an emphasis on lower-level, real-time chores. ViperOS RTOS support for the Edison’s Quark processor is currently under development, says Chase.
 

Offline johnnyfp

  • Frequent Contributor
  • **
  • Posts: 261
  • Country: nz
Re: Intel Edison
« Reply #14 on: October 08, 2014, 09:52:01 pm »
Well according to an Intel Employee on thier forum

Quote
In fact the Edison has a Quark MCU as part of its SoC but right now the Quark isn't activated. At the end of the year there is going to be a second release of the software for the Edison and, the Quark will be active.

https://communities.intel.com/message/252124#252124
 

Offline westfw

  • Super Contributor
  • ***
  • Posts: 4199
  • Country: us
Re: Intel Edison
« Reply #15 on: October 08, 2014, 10:27:34 pm »
Quote
do you really need more that 50Mhz switching speed on a GPIO?
No, but you need a lot more that the 230Hz of the v1 Galileo.  "at least as fast as an Arduino" (~4MHz) would be good.
(You have to do ... terrible things ... to get close to that 4Mhz on an Arduino.  Using digitalWrite() you are looking at less than 200KHz, IIRC.  But the point is that you CAN do such things.  On Gv1, the 230Hz pin-toggle rate was limited by the hardware design...)
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #16 on: October 09, 2014, 12:14:37 am »
A datasheet on how the Edison is wired up internally would be nice. I just can't seem to find it.
 

Offline johnnyfp

  • Frequent Contributor
  • **
  • Posts: 261
  • Country: nz
Re: Intel Edison
« Reply #17 on: October 09, 2014, 12:15:44 am »
Yeah good luck.

There are a lot of missing or not yet published documents for the Edison.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16607
  • Country: us
  • DavidH
Re: Intel Edison
« Reply #18 on: October 09, 2014, 02:35:56 am »
If truly native and the board runs at 400MHz (and I found mine running a bit over 405MHz) then it should be able to do up to 200MHz on the pins if they are all on the same address.

...

The Edison IO is restricted by the Quark processor running at 100MHz so the Edison module can drive the I/O at 50MHz tops.

Higher speed microcontrollers usually have decoupled I/O or even a decoupled I/O bus placing limitation on the toggle rate.  The earlier ARM microcontrollers did not even have interlocks so if you tried to toggle an I/O pin too quickly, the I/O became erratic.

... but in all honesty do you really need more that 50Mhz switching speed on a GPIO?
Would not the SPI or SDIO be better suited to doing anything at a higher speed?

You probably do not need 50 MHz or faster bit-bang I/O but 230 Hz is too slow and the dedicated interfaces may have better things to do.

The thing I really find annoying is if I wanted to upgrade from an 8 or 16 bit microcontroller to one of the current Intel solutions, I would still need another microcontroller (or FPGA) with a different ISA to handle low latency I/O.  If I upgrade to ARM and still need a second microcontroller, at least it could use the same instruction set and likely even the same development environment.
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #19 on: October 09, 2014, 03:16:01 am »
That's the Galileo Gen 1 that is limited by the I2C bus for GPIOs.
But in any event I don't think I'll even try to mess with the Edison at the current price offering and lack of documentation.
 
 

Online amyk

  • Super Contributor
  • ***
  • Posts: 8263
Re: Intel Edison
« Reply #20 on: October 09, 2014, 12:47:39 pm »
The amusing thing is that most if not all PC chipsets have GPIOs that can work at the base PCI frequency (33MHz); even the Galileo has 16 GPIOs in its SoC that can toggle at full speed, but for some idiotic reason they're not actually available for use.
 

Offline Scrts

  • Frequent Contributor
  • **
  • Posts: 797
  • Country: lt
Re: Intel Edison
« Reply #21 on: October 09, 2014, 02:21:05 pm »
The amusing thing is that most if not all PC chipsets have GPIOs that can work at the base PCI frequency (33MHz); even the Galileo has 16 GPIOs in its SoC that can toggle at full speed, but for some idiotic reason they're not actually available for use.

PCI required extremely tight timing. It was damn difficult to achieve that precise timing on the FPGA, so forget having it in simple microprocessors if it doesn't have dedicated PCI peripheral. On the other hand, PCI is dead. PCIe is the way to go.
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5550
  • Country: us
Re: Intel Edison
« Reply #22 on: October 09, 2014, 02:27:38 pm »
The amusing thing is that most if not all PC chipsets have GPIOs that can work at the base PCI frequency (33MHz); even the Galileo has 16 GPIOs in its SoC that can toggle at full speed, but for some idiotic reason they're not actually available for use.

Galileo Gen 2 gives native control to the GPIOs, too bad I got the earlier one :(
Also as far as I know Gen 2 is not yet supported by windows on devices, if that is the purpose for it then wait until is supported or stick with sketches.
« Last Edit: October 09, 2014, 02:29:32 pm by miguelvp »
 

Offline TunerSandwich

  • Frequent Contributor
  • **
  • Posts: 386
  • Country: us
  • I kiss on the first date
Re: Intel Edison
« Reply #23 on: October 13, 2014, 03:20:24 am »
I have been playing with my Edison for a bit now (couple weeks).  It's rubbish.  The tools for truly professional application aren't there, and the lack of documentation and object base makes it useless for "hobby types".

It's only documentation is conflicting, in error and useless.

This device was ill conceived, pushed to market too quickly and has no real application, other than novelty. 

Who cares if can run native x86, there is no IDE yet and no documentation, or object base to build from. 

Why should I have to re-invent the wheel just to use this thing?

Also I measured it's draw at 280mA (1 ish watt @ nominal cell voltage) while on the network.  Hardly usable for an embed device, running from a single 18650 Li cell, and that isn't including supporting hardware or GPIO connected hardware. 

USELESS JUNK!!!

ARM cortex A8 has broader saturation, better availability, more dev tools and far better power efficiency. 

Also why no colck REF input of V ref input?  So much for my idea about rubidium GPS sync and precision timing....

Shame on the Intel corporate team that forced the engineers to push out such a steaming pile of shit.... |O 

P.S. why in the blue hell would you build it with an internal 1.8v bus and then a 1v "precision internal V ref" and then no internal A/D.....so much UN-necessary bullshit is going to have to be waded through to properly reference this to an external 10-bit a/d.....or 12-bit or my dream of 24 bits......there goes my precision PSU/PoL/Metrics project dreams....
« Last Edit: October 13, 2014, 03:23:41 am by TunerSandwich »
In Soviet Russia, scope probes YOU.....
 

Offline SirNick

  • Frequent Contributor
  • **
  • Posts: 589
Re: Intel Edison
« Reply #24 on: October 16, 2014, 03:09:13 am »
I had no idea Windows for Devices existed.  Both that and x86 embedded seem like high-heels on a race horse to me, but OK...  While I'm skeptical to say the least, the hobbyist market is pretty resourceful and creative with new tech.  Curious to see what becomes of this.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf