Graham Pitcher

Will the universal memory always be ‘tomorrow’s technology’?

Blogs
1 min read

The development of memory technology has been a continuing theme throughout the history of the electronics industry. In the early days, RAM and ROM seemed to be enough but, since then, a variety of technologies have come. More recently, the search has focused on finding a so called 'universal memory' - something which is as cheap as DRAM, as fast as SRAM, is non volatile and doesn't wear out. The list of contenders is long and, so far, none has met the target.

One of the latest offerings is 3D XPoint technology from Intel and Micron. When unveiled in July 2015, the technology was claimed to be 1000 times faster than flash, 10 times denser than conventional memory, whilst offering the power, non volatility and cost advantages of 'all available technologies'.

But Intel and Micron didn't say how it was done. This prompted speculation that it was a reworking of phase change memory (PCM): tried and put to one side regularly during the past five decades. PCM relies on changing chalcogenide glass between its crystalline and amorphous states in order to store a 0 or a 1.

IM Flash - the joint venture between Intel and Micron - effectively said at 3D XPoint's launch that it wasn't based on PCM. Now, an executive has said the technology is based on chalcogenide materials, amongst other things.

What has become apparent is that 3D XPoint is highly complex, maybe involving 100 new materials and many more process steps. You wonder how, with such complexity, 3D XPoint can meet the cost targets expected from a universal memory, let alone the other requirements.