The Cortex -M0+ architecture is designed to provide chip-makers with the means to build microcontrollers that require “ultra low power” but are capable of 32-bit processing. Arm says it went back to the drawing board to create the new processor cores which measure 1mm by 1mm in size. It says the microcontrollers should draw around a third less energy than their predecessors, which only offered 8 and 16-bit capabilities.
If the Raspberry Pi Model B held out a little longer, imagine how much more interesting it would be.
Remember, RPi is about cost. “a little longer” might be a while yet.
How is this any way related to Raspberry Pi?
The Pi is positioned as low-cost computer, with GPU and peripherals etc.
and this M0 seems to be positioned as “high end micro-controller”.
A new Arduino however sounds like a better fit, now that they’ve announced an ARM version anyway.
Also, the news doesn’t speak anything about availability, I’m guessing it’ll be a while until devices using it are available.
I would call that: Pre-Magic Level
I mean, 32-bit, 1mm x 1mm small, and it needs very low energy. Just fascinating.
I agree. 1 mm^2 is about the size of the weave on tent canvas. That is mind-blowing.
Excuse me while my imagination pictures a “web” of Arm processors embedded between the threads…
I wonder how much power it needs, maybe if you have a tent with you could have half of the squares be photovoltaic cells.
Edited 2012-03-14 12:56 UTC
I don’t know why you would be so impressed just now, that’s not very different from the scale of ANY modern microprocessor.
OTOH, when seen in the context of scale shrink from those typical in 50s-60s, and the then-contemporary fictitious depictions of computers… but, yeah, this jump happened quite visibly some time ago.
That’s small enough to embed in something really cool, like an eyeball.
Ouch! I think I’d rather have it embedded in the surrounding tissue, and linked to the eye’s optic nerve. Having worn contact lenses throughout my life, I can say first hand that foreign objects in the eye are barely tolerable when they are custom fit, I can’t imagine even a 1mm square chip in there. Don’t forget it needs a power source too.
Then again, they are making some amazing advances with lenses containing embedded circuitry. I do agree this chip is a game changer in many ways.
http://semiaccurate.com/2011/01/28/arm-talks-about-20nm-cpus-and-ch…
I think your scale is a bit off. The M0+ is a lot smaller than 1mm x 1mm. I expect on a modern process it’s questionable if it’ll even be visible to the naked eye.
ARM’s Mali-55 GPU was 1mm x 1mm, and that was a couple of years back now.
indeed at 12k gates its more likely 1um^2 lol
1um^2 is roughly an area of a single gate (assuming a reasonably new CMOS process and not too complex gate, flip-flop’s are several times larger). But since 1mm^2 = 1e+6 * 1um^2 there is plenty of room for gates.
Even if it’s so small, so you can’t see it, core must be connected to external pins somehow.
Edited 2012-03-14 00:51 UTC
Indeed – most modern chips are pad-limited not logic-limited. With the current standard BGA fine pitch (0.4 mm) you would have at most 9 bumps on 1x1mm which is hardly useful for a 32 bit uC.
Now if they could just do this with an ARM 64-bit, quad-core, that would be really cool. Something to look forward to.
It would be nice is some chip company could put out a 23Pin ARM so I could replace the ATMega328’s that I currently use in my projects
Look at the LPC-1100 series from NXP – it sounds like something you would find useful.
Thanks for the tip, I have already used the LPC1768 and am looking at trying out the LPC11U24… But you can’t beat the form factor of the 23pin DIP for hobby work