Computronium
- A concept that can be quite misleading.
- A convenient "magical" solution for all problems regarding computation.
Contents
Required perspective
Computronium requires taking a perspective high enough that all specialized subsystems blur together into a homogeneous mass. Eventually packaged in equal unit cells that are so small that they can be treated as a continuum.
Applicability / validity of that perspective
A problem for assuming the validity of such a viewpoint is that
a big part of scale in-variance over many size scales is broken by physical scaling laws not all behaving linear with scale.
Examples where the homogeneity assumption breaks
- There are various specialized subsystems in on one chip integrated processors
- There are various specialized subsystems in on one PCB integrated computers
Examples where the homogeneity assumption holds somewhat
Within limited scale ranges:
- neuromorphic computing
- Highly regular FPGAs (field programmable gate arrays)
- highly regular memory cells (but this is already quite specialized)
- ...
Relation to the concept of "utility fog"
Just as large scale homogeneous computronium is supposed to be general purpose in that it can perform any computation. Large scale homogeneous utility fog is supposed to be general purpose in that it can emulate a wide variety of material properties. Both concepts are similar problematic in their property of brushing too many (all) details under the carpet.
Highly general utility fog has highly specialized mechanical materials as its complement.
So one might assume that computronium may have electrical metamaterials as its complement.
But it's not that simple.
(1) Computation is more than a "simple" general purpose emulation of electrical properties.
(2) In computation instead of emulation of exotic electrical properties one usually rather wants just densely packed circuits. These may stretch the concept of metamaterials.
Hypotetical "magic" materials capable of general-purpose emulation of any electrical properties could be called:
- electronium and
- electromagnetronium – when frequencies become so high that start to matter em-waves to matter
- magnetronium – just for completeness and (exploration of odd ideas)
With these one could actually place the corresponding complements like so:
- electronium <= vs => electric metamaterial
- electromagnetronium <= vs => electromagnetic metamaterial
- magnetronium <= vs => magnetic metamaterial
Even the metamaterial sides here are troublesome since:
- Electrical properties are not as independent from the chosen base material as mechanical properties. That is:
- Electrical properties much more dependent on the chosen base material as mechanical properties.
Instead of emulating many properties with just one base material. There is seems to be strong motivation to use more different base materials on a case to case redesign basis than in the case of mechanical metamaterials.
So the whole concept of electro(magnetic) metamaterials seems much less solid and well defined than the concept of mechanical metamaterials.
Note that this does snot apply to the electromagnetic metamaterials of today (2021) with way bigger microscale or macroscale meta structures. There base material and meta structuring can be perfectly separated.
