Decimal Analog Predictive Add/Subtract

[Apocryphan]

It can also stack, has anyone made a predictive add/subtract here before? I never heard of it before I was making it, then someone got on and mentioned it, and it sounded like what i was trying to make. I had to do the math the hard way to get it right.

A, B, IA(inverse of A), IB(inverse of B)

2 comparisons can turn out 4 different ways, telling us everything from when it carries to when it's a negative number.



A=2 B=7


A=5 B=4


A=4 B=5

[Apocryphan]

(2) A>B          lamps                (3)  A<B         lamps

ADD:    A-IB          2,4 carry        Invert(IA-B)     1,3
         Invert(A-B)   1,2                         A-IB       3,4 carry

SUB:     A-B           1,2                          B-A       3,4 negative 
            B-IA          2,4                          A-B       1,3 negative

        (1)A<IB          lamps               (4)  A>IB         lamps

ADD:   Invert(IA-B)    1,3                       A-IB        2,4 carry
           Invert(A-B)     1,2                       A-IB        3,4 carry

SUB:       A-B            1,2                       B-A          3,4 negative
              A-B            1,3 negative          B-IA         2,4

[optimo]

I dig it, but does it have practical use augmenting arithmetic units, or does it just standalone as a maths tool?

[Chibill]

What exactly does it do?

[Apocryphan]

The following discussion is directed to various embodiments of the invention. Although one or more of these embodiments may be preferred, the embodiments disclosed should not be interpreted, or otherwise used, as limiting the scope of the disclosure, including the claims.

a single gate that is coupled to a bus (which includes one or more signals) represents all instances of the gates that are coupled to each signal of the bus.

Read more: http://www.patentsencyclopedia.com/app/2...z4WGdCXt2Q

[Apocryphan]

I dig it, but does it have practical use augmenting arithmetic units, or does it just standalone as a maths tool?

ya, my registers will use a negative value to determine errors, so this will predict the error and correct it accordingly if i do it right.

[optimo]

ya, my registers will use a negative value to determine errors, so this will predict the error and correct it accordingly if i do it right.

that is super interesting patent docs. there might even be some new techniques to learn in there

[Magic :^)]

from the looks of the patent description, it seems like it could be a decent way to implement 3 operand addition when one value is a constant value for things like address offsets for virtual memory stuff.
the question is, does this implementation do better than MC optimized algorithms?

(also lol nice bump me)

[Chibill]

Question is there a working example of this on the server you can show me?

[LambdaPI]

Wot is dis sorcery.

[Apocryphan]

Yea,it's on my plot, I made a warp named predictive.