HP has signed up Hynix to co-develop memristor technology, now called ReRAM, Resistive Random Access Memory, and promises 10x speed gains and 10x power savings over Flash.
The deal struck between the two heavyweights aims to take ReRAM from the Lab to the Fab, where they can be mass produced and later marketed. More importantly, the companies promise to deliver in a short timeframe, with products reaching the market as early as 2013.
ReRAM is presented as a memory solution that has the potential to completely wipe Flash NAND cells off the map. Based on memristors, these non-volatile elements can change resistance when voltage is applied, only they can do it many times faster than regular NAND flash, are way more durable and will require much less power to transition states. They also are easier to stack than NAND, allowing higher density configurations.
“We believe we can create memristor ReRAM products that, at any price point, will have twice the capacity of flash memory”, said Stan Williams, HP Senior Fellow and Director of the IQS Lab at HP, describing the potential of said technology. The first devices to benefit from ReRAM will include handheld computing devices such as smartphones and media players.
HP has everything to win with this technology, if it can squeeze it out of the Hynix fabs. Hynix, however, doesn’t have all its eggs in the same basket. It has been - until now - backing another NVRAM pony: Phase Change Memory. However, the longer-term development cycle of phase-change random access memory might have tipped the balance in favour of ReRAM.
Memristors also have a second role to play. They are being hailed as the Holy Grail of ultra-tiny form factor computing devices as HP Labs demonstrated just this year that the same type of elements they had been using for storing data could also be tasked to perform logic operations, ie: a sort of ASIC and RAM device all wrapped as a single chip of silicon. The potential of this technology will remain untapped until commercial success of ReRAM is achieved.
The technology was first theorised by Prof. Leon Chua at UC Berkeley, back in 1971 and only conjured into existence in 2006 at HP Labs. Now, it seems, it is poised to become NAND Flash’s main contender, at least according to HP.