Researchers Make RAM From a Phase Change We Don't Entirely Understand (arstechnica.com)

An anonymous reader quotes a report from Ars Technica: We seem to be on the cusp of a revolution in storage. Various technologies have been demonstrated that have speed approaching that of current RAM chips but can hold on to the memory when the power shuts off -- all without the long-term degradation that flash experiences. Some of these, like phase-change memory and Intel's Optane, have even made it to market. But, so far at least, issues with price and capacity have kept them from widespread adoption. But that hasn't discouraged researchers from continuing to look for the next greatest thing. In this week's edition, a joint NIST-Purdue University team has used a material that can form atomically thin sheets to make a new form of resistance-based memory. This material can be written in nanoseconds and hold on to that memory without power. The memory appears to work via a fundamentally different mechanism from previous resistance-RAM technologies, but there's a small hitch: we're not actually sure how it works. The two mechanisms used to change the resistance have been reported in the journal Nature Materials.

Comments

Re:knowing how it works is nice, but not necessary

[morethanapapercert]

Motherboards blow 10 kinds of ass with their strict layout. Standards schmandards The nice thing about standards is that there are so many to choose from.

There are at least 12 form factors (AT, baby AT, ATX,Micro ATX BTX DTX LPX NLX Micro ATX, Mini-ITX, Nano-ITX, Pico-ITX) thatr I know of, plus numerous proprietary form factors in the SBC segment. The industry came up with those because it was useful to them to do so. Nobody forced them to and nothing stops anyone from coming up with their own form factor or proprietary extensions to existing standards. Those standards and the minimum level of interoperability are a big part of why digital technology saw such an explosion in my lifetime.

Let's look at this from both sides of the purchase: Would you want to buy a Dell tower if you knew that only Dell branded add-ons will work with it? Dell proprietary memory, Dell proprietary video card, Dell proprietary cases, Dell proprietary network cards, Dell proprietary storage and so on. We had just that situation back when I was first learning computers. With rare exceptions, hardware for a Burroughs computer had to come from Burroughs, Philips hardware had to come from Philips and so on. No single OEM aside from maybe IBM, could really achieve economies of scale, all the OEMs R&D was restricted by the need to not infringe on patented good ideas from other outfits. Shit was expensive, shit didn't work all that well, shit was difficult to adapt to user needs and it was hard to make shit talk to other shit reliably. Through sheer size, IBM managed to dominate the market and some of the early desktop standards were explicitly "IBM compatible"

On the manufacturer end, being able to use an existing hardware standard also means they are more likely to be able to use standard software implementations as well. That speeds development time and reduces R&D costs. Why re-invent the network stack, possibly introducing your own failure points, when there is already a very good, exhaustively examined and tested standard? If you make sounds cards (or these days, dedicated sound processors for inclusion in someone else's motherboard) do you really want to have to develop to meet 20 different hardware standards to match every mobo manufacturers proprietary designs, or would you prefer to just develop to the PCI standard and be able to make one device that works for almost everybody?

Finally, proprietary motherboard designs are still alive and well in the industrial/embedded segments and in laptops and other mobile devices. There the form factor is constrained by physical environment and case packaging concerns, not meeting form factor standards. What IS still being develop to standard in those markets is the interfaces. Most notebooks use the same sorts of ram, albeit with a different size and pin count, as desktop machines. They still do standard ethernet, bluetooth and so on. Also, as far as I know, Big Iron (mainframes and other very large scale computing solutions) is still largely proprietary.