@MichaelMeissner Thank you for this information. It's always interesting to hear about the ISA and amount of work that goes into a chip design.
I read recently that MicroMagic is developing a RISC-V processor capable of running at the near threshold levels: http://www.micromagic.com/news/eeNewsAnalog_Santoro_Interview.pdf
The https://www.eenewseurope.com/news/micro-magic-risc-v-processor/ is down, but this was a quote I remember saving:
"Micro Magic in Sunnyvale, California, has already claimed that its RISC-V processor design was the fastest, but in an interview with eeNews Europe Mark Santoro, CEO of Micro Magic, said the processor core had also been designed so that it can operate down to at least 350mV, near the threshold voltage of the manufacturing process."
So I know there are application processor developers out there that see the benefit of near-threshold voltage, at least for idling's power efficiency:
"And Micro Magic still isn’t revealing which foundry produced the silicon. All we are
being told is that it is a FinFET manufacturing process and that the company has compared three foundries PDKs shooting for the broadest compatibility. That indicates that
the selected process is 28nm or below and for reasons of cost and comparability is not
down below 10nm. The list of suspects is fairly narrow and includes Globalfoundries,
Samsung and TSMC, with Intel and SMIC as outside possibilities."
The most curious thing is what's absent from all of this. Back in 2011, Intel revealed a solar capable Pentium on a refined 32nm process node: https://www.engadget.com/2011-09-15...nt-near-threshold-voltage-processor-othe.html
A later analysis revealed that it uses twice the transistors (6million): https://semiaccurate.com/2012/12/20/intel-explains-claremont-the-near-threshold-solar-pentium/
"In the end, the transistor count just about doubled, the area went down by around 75x, and power shrank by about 100x at 90MHz operation. More importantly, Intel was able to run the device at near threshold voltages, and did so reliably. The design methodologies to do this are probably the hardest part of the project, and while the job isn’t finished, Intel now knows how to make a CPU run at the better part of 1GHz, then drop down to nearly zero running at near threshold voltage. I guess you could call this project a success.S|A"
The Quark processors utilized these advances, and for a time were the only x86 processors to be sold commercially at NTV, but were discontinued most likely because the market for Raspberry Pi alternatives were not as popular, profitable or powerful for general purpose applications (at least at the time), and Quark also competed with Intel's own low-end processors.
After doing some research, I found they still use NTV for IoT: https://fuse.wikichip.org/news/1119/isscc-2018-intels-self-powered-intelligent-iot-edge-mote/ The difference of course is that unlike ARM, there are no licenses one can purchase from Intel to develop a custom board, unless one is a design partner. There isn't a market yet for the need for ultra-low idling cpus that sip power like a Prius. But knowing a lot of this back in 2011-2012, I feel this project is at least trying to correct the failures of the OLPC. Btw, there are cars now that aim to get most of their power from solar: https://twitter.com/LaurinHahn/status/1362820113695662083
Around the same time that project started to get discontinued, there were almost idealistic product releases from major companies: https://news.samsung.com/global/sam...ly-solar-powered-rechargeable-netbook-nc-215s I wonder whether this was an attempt to commercialize the popularity of the OLPC charity, or whether it was just a way to look like green-washing without any serious attempt at making an ultra power efficient laptop, so well that batteries weren't needed entirely. However, I will say, that at the time, the most efficient processor sold was the Intel Atom, and used more than 5-6 watts (excluding the display). So I can see that it wasn't easy to solar power, not at least indoors. 10 yrs later, and there are many processors out there running at 5nm (Apple M1 and Samsung Galaxy .
Perhaps since the Raspberry Pi costs $5-35 for the basic boards, the power edge of 5nm just isn't there for investing in a 5nm batch at TSMC or Samsung, which I imagine has a minimum order quantity costing tens of millions of dollars. My thinking is if Reddit users were able to crowdfund a 5 second SuperBowl ad, then this thinking could be used to crowdfunding leading edge-process node for open source software and hardware designs (with the exception of EUV Foundry IP, which is both a physical and exclusive domain that FOSS- friendly foundries like Skywater will take at least a decade to reach (they are at 65nm currently).
If anyone would like to talk more about chip design and meet some of the hardware and software developers for e-paper, we have a discord server: https://forum.ei2030.org/t/join-ei-2030-discord-server/125
Thanks!
Last edited:
