Zeev's Turnips Patch-No Politics (ZEEV)

[Zeev Hed]

I doubt MRAM will become "big" in the regular DRAM business, but it has a great potential to replace various flash memory systems. If I were to compare OUM and NVEC systems, I would say NVEC has a greater chance of succeeding. As far as I know, disc drive do not use at all the spin tunneling effect. Where IBM GMR and NVEC GMR differ (both of them use the difference in tunnelling probability of electrons between two ferromagnetic layers through a thin dielectric tunneling layer, if the two layers are magnetized parallel or anti parallel to each other) , as far as can see is that NVEC sandwich its devices between additional anisotropic ferromagnetic layers, thus cross alk (magnetic) between adjacent cells is greatly reduced, since flux lines are locked within the ferromagnetic screens. That small difference has a very large impact on cell density feasible. The art prior to NVEC could not even contemplate sub micron cells, while the ferromagnetic screens now allow for cells size competing with traditional DRAM cell size. There are additional issues in the way of MRAM becoming main stream memory for "bootless" computers, but it is not inconceivable that these will be eventually solved. I believe, however, that intrinsically, MRAM will always be slower than DRAM, and thus the question if MRAM will or will not supplant DRAM. A possibiliy exists, though of a hybrid system having MRAM for operating systems only (and thus a bootless computer).