Smart pointers instead of GC?

[Ola Fosheim Grøstad" <ola.fosheim.grostad+dlang at gmail.com>]

On Wednesday, 5 February 2014 at 00:53:06 UTC, Ola Fosheim 
Grøstad wrote:
> But does D actually make sure that immutable types sit in 
> non-write-back memory marked by MTRR?

Guess I have to end this monologue by concluding that changing 
the PAT is the only reasonable way I have found so far for doing 
this. Which is too expensive for short-lived data.

So this strategy of turning immutable into non-WB memory type in 
order to avoid cache coherency latencies are only going to work 
in a OS-free environment AFAIK. (Like BareMetal OS)