Talk:Resistive random-access memory

Merge with PRAM

This is strictly the same technology as "Phase-change memory". Specific RRAM content should be merged into the "Phase-change memory" page.

Disagree: A reading of patent 6946702 suggests phase-change memory and the resistance random access memory described in the patent use a fundamentally different memory storage element. The Perovskite material requires both a positive and negative voltage to switch, which would presumably be generated by on-device charge pumps. The 1R1D device on which Winbond's patent is based is a single-write device. This forces the inclusion of reset lines to "erase" the cell so it may be reprogrammed. This appears to be Winbond's attempt to pursue something similar to Ovonics memory, so it would not be surprising to see it marketed in the same manner as PRAM and used in similar applications. However, the technology is not the same, and so I think it's fair to keep the articles separate.

That said, much of the text of the RRAM page seems to describe PRAM, not RRAM. So either the link to Winbond's patent should be removed and the page merged with PRAM, or the text describing RRAM as essentially PRAM should be corrected.Fhaigia 11:57, 15 July 2007 (UTC)[reply]

Why they shouldn't be merged There is a similarity in terminology between PRAM and RRAM but that by itself shouldn't justifying merging the articles as they stand. The biggest reason not to merge is PRAM or phase-change memory is based on measurable changes in microstructure (crystalline<->amorphous) which you can detect by cross-sectioning a PRAM cell. You won't detect the same in so-called RRAM. In fact, anyone would be hard-presssed to describe the mechanism behind RRAM, especially in perovskites.Guiding light 06:18, 17 July 2007 (UTC)[reply]

The main problem is, that for some memories it is not even clear if they are phase-change, diffusion-type or some other charge carrier effects involved. So if you want to give them a name like: RRAM or PRAM or metalization-cell you have to know the exact mechanism. By this they are merged up, because some people believe its diffusion other people believe its correlation effect of charge carriers. RRAM is just a collection of all effects not understood! Even PRAM, FRAM and MRAM are in most cases resistive... why not call diffusion-type-memory: DIFF-RAM?

Disagree: PRAM and RRAM are fundamentally different technologies. No phase change is required for RRAM, and the physics behind its operation is totally different to that of PCM. — Preceding unsigned comment added by 65.201.185.162 (talk) 16:10, 3 October 2017 (UTC)[reply]

If you are searching for a similarity and a reason to merge oxide ReRAM and PCM, then rather than looking at the composition and structure of the non volatile ( NV ) conducting states, a better place would be the post-threshold switching conducting state immediately after threshold switching. That is before the ReRAM filament grows, or for PCM the phase change occurs and the NV filaments are established.

For PCM it is fairly clear such a state does exists. In oxide ReRAMs there is a question, is the observed threshold switching transition the actual growth and link up of the NV filament. If there is an electronic conducting state for both then the idea of the suggested merge might be more appropriate. Ron Neale 86.155.110.122 (talk) 21:39, 23 November 2024 (UTC)[reply]

Companies Listed

Winbond's name is on the patent, so one assumes they are developing this technology. The technology being used at Samsung is not the same (evidenced by the fact that Winbond is not suing Samsung for infringement, nor has Samsung paid licensing fees to Winbond). Aside from the choice of name, it is unclear from the linked news article whether Sharp is pursing the technology identified in patent 6946702, or something else. I have heard Spansion is not pursuing any type of resistive nonvolatile memory, whether it be RRAM or PRAM. I am also doubtful of Fujitsu's involvement. In short, what's the source for the claim that all these companies are pursuing this technology? Added a citation tag. Fhaigia 11:57, 15 July 2007 (UTC)[reply]

I removed Fujitsu from the list since some research into their patents and commercial activities indicates that they are developing FeRAM and not RRAM.Blm19732008 (talk) 23:48, 24 July 2008 (UTC)[reply]

Fujitsu has published work on RRAM in 2007 (e.g., IEDM), although they may not have a large development program for it.218.168.209.111 (talk) 06:36, 29 July 2008 (UTC)[reply]

Literature

Elpida, Sharp et al Team for ReRAM--Sergius-eu (talk) 16:30, 14 October 2010 (UTC)[reply]

Is this something new, in develop of reram?

http://www.eetimes.com/electronics-news/4373449/UK-researchers-follow-silicon-oxide-ReRAM-route

--Palapa (talk) 14:46, 25 May 2012 (UTC)[reply]

ReRam v/s MRAM

Both are promising developments but the question remains which one's the future of memory Avieshek (talk) 08:52, 3 January 2014 (UTC)[reply]

Mechanism description

this section is quite confusing and would only be understood by someone who already understood it. — Preceding unsigned comment added by 66.212.64.252 (talk) 00:15, 14 February 2014 (UTC)[reply]

Cross-article inconsistency

What it says here:

This technology bears some similarities to conductive-bridging RAM (CBRAM), and phase-change memory (PCM).

What is says there:

programmable metallization cell

PMC is a two terminal resistive memory technology developed at Arizona State University.

"resistive memory" links to this page.

Note that—if Wikipedia is to be trusted—CBRAM is an Adesto Technologies trademarked derivative of PMC. — MaxEnt 07:13, 5 September 2017 (UTC)[reply]

NeuRRAM relevance?

"An energy-efficient chip called NeuRRAM fixes an old design flaw to run large-scale AI algorithms on smaller devices, reaching the same accuracy as digital computers, at least for applications needing only a few million bits of neural state. As NeuRRAM is an analog technology, it suffers from the same analog noise problems that plague other analog semiconductors."

This text doesn't make much sense to me. What's the relevance to RRAM, besides being part of the name? This seems like it is some kind of analog CPU, possibly using some variant of RRAM for storage.

"While this is a handicap, many neural processors do need bit-perfect state storage to do useful work."

This sentence makes even less sense in general and on this page in particular. 213.212.0.26 (talk) 13:54, 16 January 2024 (UTC)[reply]