MRAM research has been funded by the US DARPA since 1995. The efforts have culminated in the fabrication of small, working devices (e.g. 1K memories). It is now (2003) felt that commercially practical MRAMs are within reach. Expectations are that they will be significantly faster than today's DRAMs, not significantly more expensive once they come into wide use, and will store data without requiring power to do so.
Because MRAM depends on magnetism rather than current flow, contents of MRAM are preserved when power is removed. Periodic refresh is no longer required -- thus reducing memory power requirements. In portable computers, it should be possible to simply shut down the CPU retaining memory contents without battery drain. This may make "hibernation" both simpler and more reliable. It is expected that MRAM may eventually replace flash ROM memory as well as Dynamic Ram memory. It may also be possible eventually to incorporate MRAM technology into CPUs where its small cell size and low power consumption would be extremely attractive characteristics.
Current designs operate on 3volts with 15nSec access times. One initial target memory size for 2004 timeframe is 256 Megabytes So far as I can determine, MRAM devices probably will not be plug in replacements for DRAM.
(Note1: Its not entirely clear whether MRAM is an acronym for Magnetic Random Access Memory or Mangetoresistive Random Access Memory or both.)
(Note2: The term Magnetoresistive and Magnetostrictive are not interchangeable. Magnetoresistive materials change their electrical resistance in the presence of a magnetic field. Magnetostrictive materials change shape in the presence of a magnetic field.)
Return To Index Copyright 1994-2002 by Donald Kenney.