The Powder Toy

A10D28 Breakdown by Sam_Hayzen

Sam_Hayzen
Sam_Hayzen
8 / 3
16th Jan 2019
30th Aug 2020
Visual breakdown for the A10D28. Here you go, NoVIcE.

Comments

  • Sam_Hayzen
    Sam_Hayzen     17th Jan 2019
    All the hypothetical 28 1-bit adders are controlled in one instruction. Ex: While(NumB){ NumA=(NumA^NumB), NumB=(rl(NumA&NumB)) }
  • Sam_Hayzen
    Sam_Hayzen     17th Jan 2019
    And, NoVIcE, not exactly? The Logic Unit is 28 bits wide. It's kinda like... adding 28 individual 1-bit adders, taking all 28 results as a number, and taking all 28 carries and storing them as a number. Do the same with these two new numbers and repeat until there's no carry.
  • Sam_Hayzen
    Sam_Hayzen     17th Jan 2019
    That's exactly what I went for, LBP :D
  • NoVIcE
    NoVIcE     17th Jan 2019
    btw you mean a ripple carry adder which is 1 bit "wide", repeating n times to get a n bit long number?
  • NoVIcE
    NoVIcE     17th Jan 2019
    hmmm gotta research more...
  • LBPHacker
    LBPHacker     17th Jan 2019
    You don't actually need an adder for Turing-completeness, since you can emulate an adder with a Turing-complete machine. It's just that having hardware-accelerated (constant time) addition is nice :P
  • Sam_Hayzen
    Sam_Hayzen     16th Jan 2019
    Addition is done through a repetition of ANDRL and XOR. I call it "Polycarry", but it's probably already a defined method with a name; I just can't find it.
  • NoVIcE
    NoVIcE     16th Jan 2019
    I mean, maybe there is a simpler way of making a computer im not aware of?
  • NoVIcE
    NoVIcE     16th Jan 2019
    Interesting, how is there no adder yet it works?