Re: requesting support for persistent memory

[mayuresh%kathe.in@localhost <mayuresh%kathe.in@localhost>]

On Friday, June 12, 2020 02:48 PM IST, Brett Lymn <blymn%internode.on.net@localhost> wrote:

> On Wed, Jun 10, 2020 at 02:13:57PM +0200, mayuresh%kathe.in@localhost wrote:
> > i noticed; https://pmem.io/
> > they are supporting windows and linux using some 'dax' (direct-access) technology.
> > would netbsd experts too work towards bringing-in support for persistent memory?
> > i believe it could pave the way for a new breed of applications since it's a whole new programming paradigm.
> >
>
> We have RVM in pkgsrc (devel/rvm), see here:
>
> http://www.coda.cs.cmu.edu/doc/html/rvm_manual-1.html#:~:text=RVM%2C%20a%20Recoverable%20Virtual%20Memory,the%20face%20of%20system%20crashes.&text=Using%20simple%2C%20non%2Dnested%20local,the%20event%20of%20a%20crash.
>
> Not a recommendation, just pointing it out.

i was thinking more along the lines of persistent data stores, i.e. better implementations of 'rdbms' technologies.
not that the netbsd people have to "build" those applications, but providing mechanisms in the operating system to facilitate easy and efficient access to persistent memory like intel's optane (for now).
i have seen 32gb optane (in m.2 configuration from intel) go for less than us$100, and i see the industry moving towards a future where terrabyte configurations of persistent memory would be normal and affordable enough to facilitate persistent 'rdbms'-style data stores.