Mram

[DudeMiester]

I wonder how susceptible to EMP it is? It could have significant military use.

 

[_xxx_]

I wonder how susceptible to EMP it is? It could have significant military use.

I guess as much as any other piece of silicon...

 

[KimB]

I wonder how susceptible to EMP it is? It could have significant military use.

It still has wires going through it, which would quickly overheat from the EMP. I'm not sure whether or not the magnetic bits would be affected, but then again, I'm not sure it really matters.

Anyway, you can block EMP from any device with a Faraday cage. The military PC's I've seen typically have faraday cages built into the case (though admittedly it's been a long time). Damned heavy things.

 

[_xxx_]

Chal, what? Overheat? Nope. The current peak duration is too short.

To me, EMP means electro-magnetic pulse. Which produces a short, but extremely high e-m field peak. Faradey's cage doesn't help at all against the magnetic field.

Trust me, I killed many devices in the EMP testing room

 

[RussSchultz]

EMPs kill electronics by generating a potential gradient across wires, causing current flow, causing breakdown.

A faraday shield of sufficient 'fineness' should protect against EMP.

 

[Guden Oden]

I don't think MRAM would actually be all that useful for a replacement of flash memory: the density just isn't there yet

Hm, a quote in one of the earlier posts claimed 'similar' density to DRAM, and that is much denser than flash.

and neither is there much need for the added performance in such a format.

Uh, why not, pray tell? Flash is dreadfully slow by contemporary standards. Only on large block reads/writes do some solutions start to reach acceptable levels, on smaller accesses speed is almost pedestrian. A faster alternative would be much welcomed I'm sure, I don't think nobody ever said in the history of computing, "nah. This is fast enough, we don't need anything better, ever."

Besides, let's not forget the 'unlimited' writes capability. Flash simply can't be trusted in the long run due to its proneness to breaking down. You can't tell when it'll happen, just that it WILL happen, sooner or later. Very soon, in the case of poor quality flash memories, which again you can't tell from the outside.

Sure it is expensive NOW, but as production volume rises, prices will plummet. Do you remember when harddrives cost several bucks per megabyte? I do. And look at prices today.

 

[mczak]

Hm, a quote in one of the earlier posts claimed 'similar' density to DRAM, and that is much denser than flash.

No, this is not true. Nowadays (at least nand) flash has a smaller cell size than dram, and for some reason it looks like it is easier to make it smaller than with dram. A quick googling didn't turn up directly comparable numbers on comparable process technology, unfortunately.

 

[Squeak]

If they can get the density up (by a lot) and the price down (ditto (and I don't see why not with either)) then MRAM would be the solution to all of the current problems of memory latency and fragile HDDs .
It's basically the return of core memory, only now using litographi instead of underpaid Asians to tread the magnetic beads/bits.

 

[hoom]

Apologies for the Necro (time flies I thought this was an early 2007 thread) but has anyone heard any further info about how MRAM is coming along?
Like when will I be able to buy the stuff

 

[Blazkowicz]

imagine 4K of mram in a micro-controller, now that might be useful.

 

[pcchen]

The last news of MRAM I know is about a new NEC chip which runs at 250MHz. The capacity is 1Mbit per chip.