0

How does the technology keep improving despite having everything discovered already? I mean the same sized chips and electronics are used from year to year but with every new version of the main board or something like that the parameters are better. The humanity already knows all the chemical details and all the possible combinations but how are microchips still improving? Why they can't produce the best possible chipset already?

7
  • 6
    "having everything discovered already", "The humanity already knows all the chemical details and all the possible combinations" - you are very, very wrong on both accounts. But anyway, even if we perfectly knew the fundamentals, going from a quantum mechanical theory to predictions of chemical reactions is highly nontrivial - the equations are extremely complex, in general not solvable analytically and the parameter space is so huge that any computer simulation is impossible for anything beyond some simple atoms. Nov 15 '21 at 10:46
  • So basically you are telling that even if we assume we know all the chemical combinations, we still won't be able to predict how they will interact in different scenarios and there are plenty of space to be discovered as there is no computer that can solve anything related to atom level Nov 15 '21 at 10:51
  • 1
    It wouldn't surprise me if most, if not all, chemist would disagree that "the humanity already knows all the chemical details and all the possible combinations".
    – UncleBod
    Nov 15 '21 at 11:29
  • 1
    The chips are not the same size.
    – user253751
    Nov 15 '21 at 11:52
  • 2
    Some performance improvement comes from algorithmic improvements in the chip logic. I don't know any recent examples, but pipelining, speculative execution, branch prediction, etc., come to mind. None of this relates to "chemistry". Nov 15 '21 at 12:23
2

I don't know where you came up with the idea that 'the same sized chips and electronics are used', because that's not the case. Take a look at Wikipedia's 'List of semiconductor scale examples'. This lists the size of various features and parameters used in integrated circuit manufacturing over time. Broadly speaking, the Arm chips in a smartphone from 2017 have transistors 1% of the size (and 0.01% the area) of a CPU like an 80386 from 1987.

There are some questions on RCSE "When did MOS Technology upgrade to 5µm?" - "Why did the VIC-II and SID use 6 µm technology in the era of 3 µm and 1.5 µm?" specifically relating to electronic process size. This improvement never stopped, in the CPU space, the GPU space or any other space. "Intel's Alchemist graphics cards will be built on TSMC's 6nm process node" from August 2021, for example.

Irrespective of the likely apocryphal quote that "everything that can be invented has been invented.", everything has indeed not been invented.

Materials scientists experiment with new combinations of alloys, dopants and so on on a continuous basis. Technological research has not, and has never, stopped. Some current fields that spring to mind relevant to (retro)computing: new types of semiconductors, new methods of manufacturing, more efficient - less leakage of current/power/heat, more stable - higher yields, more reliable under environmental changes: temperature/shock/humidity/vacuum, smaller, faster, miniaturisation of sensors (MEMS).

If you want to read some of the absolute cutting edge of (publically released) science, go to arkiv and read the abstracts of some preprints that interest you (these are scientific research reports that have been/are soon to be submitted for professional evaluation before publication in a journal). There's some there about chemical simulation, if that's what interests you. And even if all were known about the interactions between various materials - the simulations themselves would be subject to continuous improvement, bringing more complex and intricate systems within our reach.

For an example in industry, I picked a phrase off the top of my head, 'new semiconductor substrate', and found the following press release.

1
  • Definitely correct me if I'm wrong on any of the figures!
    – knol
    Nov 15 '21 at 11:10

Not the answer you're looking for? Browse other questions tagged or ask your own question.