Jump to content
Sign in to follow this  
Guest_Jim_*

Potential Universal Memory for Computers Developed

Recommended Posts

Modern computers and devices have can have a variety of ways to store data, in part because each memory type has its advantages and disadvantages. For example, DRAM can be written to and read from very quickly, and it does not take much energy to operate either, but is volatile, so the data it holds needs to be refreshed or it will be lost. Flash memory is non-volatile, so the data it holds does not need to be refreshed to survive for long periods of time, but it also takes substantially more energy than DRAM to operate. Researchers have been working on creating a universal memory for decades, one that offers high speed, long data retention, and achieves this at low energies, and now those at Lancaster University have patented what may just be that solution.

This new memory type, which has been patented in the US and has another patent pending, uses an oxide-free, floating-gate memory cell based on a compound semiconductor. The compound semiconductors used allow its switching voltage to be less than 2.6 V and for data to be preserved for very long times. Potentially data could be stored indefinitely, though that might not be realized in a practical device. The amount of energy it takes to switch states in the memory decreases as the device is shrunk, such that on a 20 nm node it should take 100 times less energy than DRAM and 1000 times less than Flash.

Already a number of companies have expressed interest in this technology or were actively involved in the research itself, so perhaps we will see this memory solution coming to practical devices in the near future. As the DRAM market is on the order of a $100 billion market, there is definitely going to be an interest in related and competing technologies.

Source: Lancaster University and Nature



Back to original news post

Share this post


Link to post
Share on other sites
when thinking about artificial intelligence and sentient electronica, this just makes total sense, if only in terms of cognitive appraisal... an AI like a human needs the ability to store learning and experience, while at the same time cope with new stressors and modify accordingly. switching from non-volatile to volatile is what we humans do...

Share this post


Link to post
Share on other sites
Sign in to follow this  

×