Posted on 07/17/2012 10:06:31 AM PDT by Ernest_at_the_Beach
Memory maker Micron has become the first firm to put Phase Change Memory (PCM) - one of the possible non-volatile successors to Flash - into mass production. Caveat: we're talking the technology's use for mobile devices only.
Micron is pitching its initial offering at mobile devices, combining 1Gb of 45nm PCM and 512Mb of low-power DDR 2 memory into a single, 1.8V chip, though both types of memory are implemented on separate dies. Cramming memory and storage into a single chip will appeal to mobile device developers who generally favour a high level of integration, the better to get more into small, phone-size cases.
PCM uses heat generated by an electrical current to flip a material between separate physical states. Each of these phases has distinct properties that can be sampled, typically by measuring its electrical resistance, allowing each cell to represent one or more bits.
Unlike Flash, PCM cells are erased at the bit level, not in entire blocks, so the writing process is much faster than that of Flash chips. PCM's write latency is typically 1µs - two orders of magnitude less than Flash.
Read latency, says Micron, is in the 50-100ns range - slower than Ram, but considerably faster than Flash's 10-25µs (10,000-25,000ns).
PCM cells are more resistant to high-energy radiation, making it suitable for space applications for which radiation-sensitive Flash is inappropriate.
Like Flash, PCM has limited longevity, but the technology's proponents claim it operates over many more millions of write cycles than Flash.
Micron said its PCM delivers a random read performance of "up to 400MB/s" and an endurance of "more than 100,000 write cycles". Those stats aren't significantly beyond modern Flash specs, but then these are very early days for commercial PCM. ®
Technically, yes, but it won't be competing with flash for storage volume any time soon. It may be useful where small amounts of very fast storage are needed, but you won't see it in a phone or tablet for years at least. Sad, because the speed of the storage is becoming a big obstacle to modern devices. That's probably why Apple bought that Israeli NAND optimization company.
I guess its commercialization will parallel what we have
seen with other types of memory. There will likely be
found a sweet spot in the array of devices out there where
its price/performance characteristics make sense for the
given application. Obviously as time goes on economies
of scale will kick in and prices will drop. Early adopters
always pay a premium. Later adopters get to back up the
truck and load up. It’s always seemed to me that the
rate at which prices drop and production ramps up is faster
than initially expected but fundamentally the technology
has to deliver a clear advantage for that to happen. It
would seem that that is the case here but I’m no expert
and time will tell.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.