anon (spam.delete.delete@this.this.spam.com) on December 1, 2018 8:21 am wrote:
> Michael S (already5chosen.delete@this.yahoo.com) on December 1, 2018 7:56 am wrote:
> > Howard Chu (hyc.delete@this.symas.com) on December 1, 2018 5:23 am wrote:
> > > Personally I don't believe Optane DIMMs are going to replace DRAM, but tech
> > > like STT-MRAM has a good shot, if they ever reach parity to DRAM density.
> >
> > I don't quite get it. Why parity is enough? Doesn't new tech has to be at least
> > 1.5x to 2x denser in order to get a shot at replacement of old tech? Especially
> > something like MRAM that seems to get an inherent disadvantage in write energy.
> >
> > Or do you consider non-volatility an advantage?
> > I used to think about non-volatility as minor disadvantage, but in lights of all those new
> > security threats around I am starting to think that may be disadvantage is not so minor.

What security threats are you thinking of? I believe with non-volatility and
non-destructive reads, MRAM is immune to rowhammer-style attacks. (But I don't
have any references that claim to have tested or verified this.)

Greater density would be a nice bonus, but yes, I believe reaching parity
is enough, because non-volatility is a significant advantage in terms of
system reliability, performance, and power savings. Eliminating the DRAM refresh
cycle and its periodic bottleneck is a significant win.

It's nice that Optane DIMMs are denser than DRAM, but they're still far behind
in latency. Nobody is going to use a machine whose system RAM is 100% Optane,
it would be much slower than a DRAM-based machine.

> For a database it's an advantage because then you don't need to guarantee
> that the RAM can't lose power before all data has been written to disk.
> For everything else you're probably right.

For any application that ever needs to store persistent state - which is
pretty much all of them. Would let you avoid lengthy fsck procedures on
most filesystems.