I say what I mean, and mean what I say.

Which is different than comprehending what you say...

Perhaps you are unaware that there were such things as PMOS Logic and NMOS Logic of which actual LSI semiconductors were built. This is the context in which I was writing. Both technologies are now thoroughly obsolete.

I have PMOS and NMOS logic chips in my inventory, so it's safe for me to say that I'm aware of both technologies and that I've used them in my designs. You may call them obsolete, but they still exist in every CMOS device and we couldn't have CMOS without them.

Heck, I even personally know some of the people who designed some of the NMOS parts I use...

 

Intel is a great company that made the industry possible in the PNW but Intel just no longer makes the products needed for IOT embedded applications at the tip of the spear.

I wouldn't bet against Intel on IoT; they were one of the companies that launched it and have a strong desire to be successful in that market. How many companies could absorb a $1B annual loss in a market (mobile) and still have the internal fortitude to continue?

 

I wouldn't bet against Intel on IoT; they were one of the companies that launched it and have a strong desire to be successful in that market. How many companies could absorb a $1B annual loss in a market (mobile) and still have the internal fortitude to continue?

They will not make the chips on the "things", too expensive and over engineered for the single (or few) simple tasks that the "thing" must do (lamp: on/off, lock:locked/open, garage door: up/down, ...). Definitely over engineered when a uC can handle it.

 

I wouldn't bet against Intel on IoT; they were one of the companies that launched it and have a strong desire to be successful in that market. How many companies could absorb a $1B annual loss in a market (mobile) and still have the internal fortitude to continue?

Not many but that doesn't give me much faith they can win another sector with things like Edison or Galileo (that are brilliant engineering but not embedded products) by dumping billions into it when the 8 bit companies are making hundreds of million using archaic equipment, are flush with cash and are upgrading equipment from Intel's hand-me downs to be only 3 generations behind. Intel is a needed component in the total picture but you need real 5 volt logic that runs from DC to max clock speed with ultra low leakage that can take real-world EMI/ESD with simple I/O tasked based programming sets.

 

Now that it's public you can see how companies are building the capability to make the IoT possible at the embedded level.
http://www.micrel.com/
http://www.micrel.com/_Foundry/Micrel_Foundry_Brochure.pdf

What a company like this has at the Fab level is so far below what's possible at the 'End of Moore's law' its almost laughable to the Intel's of the world but it's all that's needed to make the parts the big dogs need to talk to the world.

"They own the clouds but we own the dirt"

 

Here's an interesting article relevant to this thread:

http://www.cnet.com/news/life-after...dclid=1-1873234-8703601-117639148-290528809-0

In the end it mentions another possibility that hasn't been mentioned here: spintronics.

 

Metal spintronics is quite advanced with most of the research and has actual products based on it. Semiconductor spintronics (nonmagnetic materials) is still really at the 'Paper' stage. Transistors are easy, interconnects are the big bottleneck and most of the new technologies do nothing to solve that problem.

 

As suspected, the reports of the death of Moore's Law have been greatly exaggerated:

http://www.nytimes.com/2015/07/09/t...hips-more-powerful-than-any-in-existence.html

 

As suspected, the reports of the death of Moore's Law have been greatly exaggerated:

http://www.nytimes.com/2015/07/09/t...hips-more-powerful-than-any-in-existence.html

"As points of comparison to the size of the seven-nanometer transistors, a strand of DNA is about 2.5 nanometers in diameter and a red blood cell is roughly 7,500 nanometers in diameter."

WOW ...

 

As suspected, the reports of the death of Moore's Law have been greatly exaggerated:

http://www.nytimes.com/2015/07/09/t...hips-more-powerful-than-any-in-existence.html

The bottleneck is still EUV source power. The current technology using pumped lasers needs the stepper optical path to be in high vacuum and the best power sources are still under 100W.
http://www.nanowerk.com/spotlight/spotid=35314.php
http://www.eetimes.com/document.asp?doc_id=1322626

 

The bottleneck is still EUV source power. The current technology using pumped lasers needs the stepper optical path to be in high vacuum and the best power sources are still under 100W.

EUV is constantly missing commits. The light source is 13nm; it needed to be available for 14nm. By the time it's available, double EUV patterning will be required. The wimpy power source will require longish exposure times which will kill throughput.

 

a lot of stuff was invented here in Wichita, the old NCR plant had quite a few things, like the cd rom, and others. in conjunction with their silicon on saphire plant in Colorado Springs, they made quite a few interesting technologies. All I did was test and repair boards for the systems they were making, so I wasnt involved, but they also came up with SCSI and the controller chips for it.
Moores law isnt obsolete, its just waiting on the next (whatever that might be) breakthrough to restart.

 

2021 now:

https://www.extremetech.com/extreme...dead-by-2021-to-be-replaced-by-3d-integration

 

Link copied from http://forum.allaboutcircuits.com/threads/intel-finds-moores-laws-next-step-at-10-nanometers.130856/

http://spectrum.ieee.org/semiconductors/devices/intel-finds-moores-laws-next-step-at-10-nanometers

 

https://www.fastcompany.com/40401818/moores-law-survives-but-its-not-what-it-used-to-be

For Intel, it was an important announcement: Moore’s law is not dead.
...
“Number one, too many people have been writing about the end of Moore’s law, and we have to correct that misimpression,” Mark Bohr, Intel’s technology and manufacturing group senior fellow and director of process architecture and integration, says in an interview. “And number two, Intel has developed some pretty compelling technologies … that not only prove that Moore’s law is still alive, but that it’s going to continue to provide the best benefits of density, cost performance, and power.”

 

https://www.theregister.co.uk/2017/06/20/intel_joules_edison_galileo/

Intel has discontinued three of its offerings for the Internet of Things and embedded device markets.

The chipmaker said in a series of low-key product updates that it would be killing off the Edison , Galileo and Joule compute modules and boards over the second half of the year.

The notices mark an ignoble end for three lines that were once seen as key to Chipzilla's IoT and connected appliance strategies.

https://www.electronicsweekly.com/news/business/intel-cut-iot-jobs-2017-07/

Intel being Intel again.

 

The end is nigh!

 

The end is nigh!

Plenty of money to be made off the bleeding-edge.

https://www.eetimes.com/document.asp?doc_id=1333637

“The lion’s share of our customers…have no plans for” 7nm chips. Industry-wide demand for the 14/16 node was half the volume of 28nm, and 7nm demand may be half the level of the 14/16nm node, Caulfield said.

“When we look out to 2022, two-thirds of the foundry market will be in nodes at 12nm and above, so it’s not like we are conceding a big part of this market,” he added.

 

When enthusiast microprocessors are around a dozen cores and many families have a dozen or so compute devices (including tablets and smartphones), who really needs microprocessors with smaller geometries?

Global Foundries just announced that it was throwing in the towel at 7nm. Intel was late with 10nm. The design rules are too complex for any but the hardcore layout designers. EUV was such a disappointment that I stopped following its progress...

I'm satisfied with the compute power from computers that are pushing their 10th birthday and my daily driver is a Pentium4 running XP Pro; it's fine for email, word processing, image editing, and schematic/PCB design. It's soon to be replaced by a Win10 computer only because of security issues.

With 4 laptops, 5 tablets, 4 smartphones, and half a desktops, enough is enough.

 

And then of course stuff like this keeps popping up -

http://fortune.com/2018/08/18/germa...-quantum-transistor-using-just-a-single-atom/

I worked as a production EE on a PMOS line in mid 1970's, built one of the first big memories
for DOD, a 256K bit serial shift register. +5/-12 power, clock lines had more capacitance than
the earth*). Drum replacement stuff. Yield started at 1/2 die /wafer (we had to run 2 wafers to get
one good die). When I transferred it into production we were getting 5 die / wafer.

Is the Planck length the ultimate geometry limit ?

Regards, Dana.

* What is the effective C between the sun and earth ? Maybe that's where all the noise is coming
from we fight each day in designs.