News New 'DRAM+' memory designed to provide DRAM performance with SSD-like storage capabilities, uses FeRAM tech

[Admin]

FMC and Neumonda develop non-volatile FeRAM products, revive memory production in Germany.

New 'DRAM+' memory designed to provide DRAM performance with SSD-like storage capabilities, uses FeRAM tech : Read more

 

[A Stoner]

Germany does not create much power any more. I sure hope the process of manufacturing does not need much electricity.

 

[Hooda Thunkett]

I'm sure we all feel like we've seen this news before. Names change, claims don't.

 

[usertests]

I'm sure we all feel like we've seen this news before. Names change, claims don't.

Let them cook (us up some universal memory).

 

[joeer77]

I've been working with HfO2 for 21 years. I've never heard of a ferroelectric effect for HfO2. It is a high-k insulator used for gate oxide (insulator) or as a capacitor dielectric.

 

[m3city]

Goodram makes ssd, ddr, usb drives. That is actually the last factory in europe/poland og such kind.

 

[3en88]

Since it's German, if their claims hold true, one should expect atrocious quality and exorbitant prices from their products despite massive public funding and subsidies on every level, local, federal and EU.

 

[JRStern]

Doesn't "ferro" mean iron?
SMH
Yes I see the term is already established as meaning roughly "magnetic like iron", like ferrite cores.

 

[bit_user]

So, is it susceptible to external magnetic interference? Does that mean the chips would need to be encased in a steel enclosure, like HDDs?

 

[kerberos_20]

I've been working with HfO2 for 21 years. I've never heard of a ferroelectric effect for HfO2. It is a high-k insulator used for gate oxide (insulator) or as a capacitor dielectric.

https://www.nature.com/articles/s41524-024-01352-0

 

[rdgordon]

How is this different (in function) from Optane pmem? Something released almost a decade ago, before discontinuation due to a lack of interest?

 

[usertests]

How is this different (in function) from Optane pmem? Something released almost a decade ago, before discontinuation due to a lack of interest?

Optane had worse speed and latency than DRAM by around 10x. It also did not have the "infinite" write endurance of DRAM so it wasn't reliable enough to act as a complete replacement for it. It had better density and $/GB than DRAM, but it lagged behind NAND. NAND had experienced a scaling crisis that was solved by the move to 3D NAND, allowing it to become much denser and cheaper.

So Optane ended up occupying a tier in between DRAM and NAND. It was very useful for some customers, and might be considered a "perfect" boot drive for consumers, but it was apparently not enough of a market success for either Intel or Micron to stay the course.

Some enterprise/industrial applications call for non-volatile memory, even if it's expensive and in small amounts. DRAM+ seems to have improved scaling/density than previous non-volatile memories. That's about all we can say for now.

If we want "universal memory" in the future, we would want something that has all the good properties of NAND (cheap, dense, nonvolatile) and DRAM (fast, low latency, high endurance). This could replace either technology in many applications, and unify storage and memory which could speed things up tremendously. The ultimate computer might be an APU with terabytes of universal memory on package as an L4 cache, because reducing the distance between components reduces latency.

If universal memory remains vaporware, it's business as usual for us. But at least we might see the commercialization of 3D DRAM in the 2030s, for density and cost improvements.

 

[bit_user]

but it was apparently not enough of a market success for either Intel or Micron to stay the course.

Here, I have to point out that Intel pulled some seemingly sketchy moves to basically sabotage Micron's attempts to deliver its own 3D XPoint-based products. Apparently, the partnership only went as far as letting Micron fab the chips, as Intel essentially claimed that Micron lacked a share in the IP needed to deliver finished products (as best I understand it).

I did a quick web search to find related articles, but instead happened upon a tale of intrigue around the technology I never knew existed. Not only was Intel suing Micron to stop it taking IP it claimed wasn't theirs - and without which, they couldn't deliver products based on the technology, but it seems Intel and Micron were also sued for infringing the IP rights of the tech's original developer. There were a few other lawsuits, involving individuals, I ran across, that I'm not sure were related to either of the above.

3D XPoint patent suit against Micron and Intel is allowed to proceed – Blocks and Files

A US lawsuit filed two years ago by the liquidator of a bankrupt patent company against Intel and Micron has been allowed to proceed. The ruling issued on October 1 by Judge Thomas J. Tucker, exposes Intel to claims that its has no right to its core 3D XPoint technology. ECDL Trust accuses...

blocksandfiles.com

Anyway, so I just wanted to point out that Micron's lack of follow-up wasn't necessarily due to anything but Intel's backstabbing. However, Intel's discontinuation of Optane is another matter and one I think you covered well.