4 ms·

This is sad to see. I was excited by the idea that 3D XPoint research might put us on the path towards memristors, but it seems like the tech is likely too "ear

This is sad to see. I was excited by the idea that 3D XPoint research might put us on the path towards memristors, but it seems like the tech is likely too "early" to easily sell. By that I mean: Computers are built around the idea of RAM being volatile. Having it stay populated on reboot is weird -- that's the job of the disk today. Writing applications that have one shared, unified space for storage and memory isn't something that any tool chain does today (afaik). There is too much legacy and not enough incentive to lean in on this. I hope somebody figures out how to take this tech to production. I'd love if my phone weren't chewing battery keeping DRAM alive all the time.

jiggawatts 4y ago
As a potential customer I looked into these, but products were few and far between. More importantly, they weren’t competitive with ordinary DRAM for either capacity or price. The promise was that we would be plugging terabyte sized XPoint DIMMs into ordinary servers and getting massive capacity benefits at a small performance hit compared to DIMMs. In practice the sizes available weren’t anywhere near that large and cost more. So… what’s the point?