8 comments

[ 0.17 ms ] story [ 13.3 ms ] thread
My question is: Why didn't Intel already know about this? They have more real and theoretical knowledge of the behavior of silicon and various silicon-based materials than anyone else.
Their research labs have quantum computers, I'm sure they are well aware. As for pushing this to market, no hurry, they're in the business of selling silicone 8086 based processors, they'll probably keep doing that for as long as possible.
The article changes tenses a lot, which makes it almost impossible to tell if this has been created already, or is waiting to be created. Dr Kenyon's site doesn't really help in that matter.
Wow, only 3,000 cycles. I haven't heard about this property of memristors before. I guess they still have a long time to go before becoming production-ready. On the good side, they look very, very dense, so maybe some kind of ECC would help?http://www.ee.ucl.ac.uk/~tkenyon/Photonic_Materials/Silicon_...
Well, IMEC was able to create ReRAM cells with a billion read/write cycles, and HP has talked about accumulating data at a trillion cycles, so hopefully the commercial product will be much more durable.
No, it isn't "as cheap as chips."

I've looked, and I can't buy it at any price, which means it's basically infinity. (Searching Newegg returns this: 'We have found 0 items that match "memristor".') However, lots of places are selling all kinds of things with conventional RAM in them, for prices a whole lot lower than infinity. (Unfortunately, hot air is not included.)