Unfortunately I missed the European showing of Panasonic's microcontrollers with on-chip resistive RAM at this year's Embedded World exhibition.
The news that Panasonic was offering two types of 8-bit microcontroller – low power and general purpose – broke in May 2012 and the good news is that to accompany the appearance at Embedded World Panasonic made an English language Youtube video to explain the technology.
The move marks an alternative to the MCU plus non-volatile ferroelectric memory offering being made by some companies.
The resisitive RAM is based on tantalum oxide layers between tungsten plugs in a manner similar to a number of metal-oxide ReRAM schemes that are being researched by numerous companies. However, Panasonic has gone further by integrating the memory in a logic process and offering the resulting microcontrollers in evaluation kits.
The microcontrollers come with 128-kbytes of non-volatile instruction memory and 8-kbytes of non-volatile data memory. The low-power version operates at down to 0.9-V where it runs at 64-kHz clock frequency. The general-purpose version can be selected to run at between 1.8-V and 3.6-V at a clock frequency of 10-MHz.
The main claim for the embedded non-volatile memory is that it provides 10 year retention and good endurance of at least 100,000 program erase cycles at lower supply voltage and faster programming times than competitor memories, which results in lower power consumption.
Thanks for your comment on FRAM MCU (Wolverine).
However, I am not sure about ReRAM ALWAYS having limited write cycling. Some of the learned papers on metal-oxide ReRAM have shown extremely high cycling endurance.
Samsung and KAIST reported on a tantalum-oxide bilayer structure (same as Panasonic) with 10^12 cycles endurance. http://www.eetimes.com/electronics-news/4217799/Samsung-trillion-cycle-ReRAM
Good for the start, but too early for cheer. From years ago Fujitsu and TI is promoting FRAM MCU, non-volatile RAM, which reads and writes data at 1.8V with low power consumption. Actually Panasonic also has a very talented but too closed FRAM team beside ReRAM. Panasonic still needs much more time to be able to stably manufacture and be accepted by the market. You can write infinite with FRAM, But ReRAM will always have limited write cycle(100K) which is extremely unuseful for frequent logging applications. They say smart meters as target application in the video, with 100K write, forget it.
My apologies. Sometimes your valued contributions are overlooked.
Noticed in your write-up the last part that you commented on:
"..,full compatibility with standard manufacturing back-end-of-line flows, including robust data retention through the solder reflow process." said Narbeh Derhacobian, CEO of Adesto, in the same statement."
A broadside on Phase Change from Narbeh - the high temperature experienced during reflow will erase it...,
If you mean this announcement
from April 2, then I did not.
If you mean something else, then possibly I did.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.