Intel's new microcontroller

[autobot]

Here's the datasheet for intel's new microcontroller , supposed to cost $2-3 depedning on volume, with price decrease in the future.

datasheet

http://www.intel.com/content/www/us/en/embedded/products/quark/mcu-d1000/quark-d1000-datasheet.html

nothing special, right ?

[c4757p]

It looks about as capable as a lower-end M0 and costs 2-3x as much. I don't understand who is supposed to want this.

[asgard20032]

And it don't get lot of peripheral

[ralphd]

Only 33Mhz.  Many of the cheap M0s go to 48Mzh.

[ataradov]

That 13 cent Cortex-M0 device that just floated in another thread got more going for it than this chip from Intel. The purpose of this device is not clear.

[richardman]

The purpose is for Intel to get to the IoT market which is suppose to be huge and every silicon vendor wants a piece of that pie. Intel has 2 major advantages: their own fab and their own instruction set architecture. Unfortunately, neither of which will help them much in the embedded space. MCUs do not push the envelope of fab manufacturing and the ISA only helps if you are running Windows.

OTOH, this is a market that Intel cannot and will not ignore, so it would be interesting to see what they will do.

[ataradov]

OTOH, this is a market that Intel cannot and will not ignore, so it would be interesting to see what they will do.

Give up and license ARM core. They already use AMBA for interconnect, so it will be a drop in replacement. And after that they need to work on peripherals, right now there is nothing to look at there.

All of this can be achieved by acquiring some MCU manufacturer. Everything seems to be for sale these days.

[coppice]

The purpose is for Intel to get to the IoT market which is suppose to be huge and every silicon vendor wants a piece of that pie. Intel has 2 major advantages: their own fab and their own instruction set architecture. Unfortunately, neither of which will help them much in the embedded space. MCUs do not push the envelope of fab manufacturing and the ISA only helps if you are running Windows.

OTOH, this is a market that Intel cannot and will not ignore, so it would be interesting to see what they will do.

Why can't they ignore it? They used to be a substantial player in the MCU business. First they dropped the low end (8051), and later the higher end (80960, etc). Now things are much more aggressive in the MCU market than when Intel left. What would make sense for Intel now which didn't make sense then?

[LabSpokane]

The real winners in the IoT lottery will be those who don't buy a ticket. 

[Mechatrommer]

I don't understand who is supposed to want this.

intel's fans, worshippers and nuts.

[c4757p]

Fanboys don't sustain a product line, though.

[daqq]

The ADC parameters and the built in step down converter (just add coil and diode) are nice, but aside from that I can't see what would motivate me to prefer this to, say, something from ST or similar.

[andersm]

Now things are much more aggressive in the MCU market than when Intel left. What would make sense for Intel now which didn't make sense then?

There's much more money there now, and they can make it run x86.

[c4757p]

Intel has 2 major advantages: ... their own instruction set architecture.

Why is this an advantage? If I were evaluating someone's microcontrollers for use in a product, the use of a nonstandard ISA only implemented by one company with a very small microcontroller inventory would almost 100% disqualify them.

[richardman]

Intel has 2 major advantages: ... their own instruction set architecture.

Why is this an advantage? If I were evaluating someone's microcontrollers for use in a product, the use of a nonstandard ISA only implemented by one company with a very small microcontroller inventory would almost 100% disqualify them.

It goes with the 2nd part - the x86 ISA is the reason why Wintel dominates the PC world. Windows had been ported to DEC Alpha, MIPS, PPC, 960, ... all dead. When you throw enough smart people at a pig, eventually it can look like a swan. The obvious point is that the x86 ISA does not carry the same advantages in the embedded world, unless of course if everything runs Windows 10.

[JoeN]

More information.  Registers, memory map.

http://www.intel.com/content/dam/www/public/us/en/documents/guides/quark-d1000-user-guide.pdf

More information, download the IDE:

https://firmware.intel.com/projects/quark/intel-quark-microcontroller-d1000

Buy some:

http://www.mouser.com/search/ProductDetail.aspx?qs=MNIFWdyJgbqv4nzs%252bL2%2f4g%3d%3d

[Scrts]

That 13 cent Cortex-M0 device that just floated in another thread got more going for it than this chip from Intel. The purpose of this device is not clear.

Sorry missed that. Can you post a link to that thread?

Thank you.

[ataradov]

Sorry missed that. Can you post a link to that thread?

https://www.eevblog.com/forum/microcontrollers/new-st-nucleo-boards/msg791932/#msg791932

[ataradov]

More information.  Registers, memory map.

At least documentation seems to be reasonably complete.

Now where can we get header files for this?

It would be an interesting toy if someone came up with a cheap eval board. Beyond that, I rally don't see how this can be successful.

[tszaboo]

The ADC parameters and the built in step down converter (just add coil and diode) are nice, but aside from that I can't see what would motivate me to prefer this to, say, something from ST or similar.

By the looks of it, they spent some time on the analog peripherals.
I think at least it shows that they are going to the right direction. When they targeted the IoT with 400 Mhz x86 with DDR3, doing a wireless temperature measurement, I was laughing my ass off.
We dont see the roadmap, probably we get micros with built in wireless, usb, and stuff.

[legacy]

thank you, I will notify it to my executive manager, in order to get a few samples, it "might" be interesting for us.

[mikerj]

Looks like a decent enough device; very low power sleep, nice fast ADC, loads of comparators, 32 bit timers etc. but they are very late to the party for a device in this class, and if the price estimates are correct it's way too expensive.

[BloodyCactus]

interesting.  32bit x86 uc... will have to see where it goes.

[MT]

It goes with the 2nd part - the x86 ISA is the reason why Wintel dominates the PC world. Windows had been ported to DEC Alpha, MIPS, PPC, 960, ... all dead. When you throw enough smart people at a pig, eventually it can look like a swan. The obvious point is that the x86 ISA does not carry the same advantages in the embedded world, unless of course if everything runs Windows 10.

Pig's are often quite calm animals while swan's is actually very aggressive. Win10 needs loads of speed to be useable, 33Mhz wont cut.
286,386, and even low power PC on a chip 486 have been made by multiple vendors in the past all for embedded use.

[BloodyCactus]

well, wont be running second reality or panic on this thing but hmm i am curious about its tooling and debugging...