Arithmetic Using Simulated Qubits - Brain Theory from a Circuits and Systems Perspective - page 124

Divide the number n by integers m (beginning from 2 up to the square root of the
number) in a search for exact divisors. All division can be in parallel. One or more
of the remainders would be zero for a non-prime number; so one can know that a
given number is not a prime. But if all remainders are finite, one can know it is
a prime number.
This routine can be implemented more efficiently if a complete list of primes up
to the square root of the number is available and is loaded into its respective
register. Then trial divisions need to be checked only for those m that are prime.
For example, to check the primality of 37, only three divisions are necessary
(m ¼ 2, 3, and 5), given that 4 and 6 are non-prime (Wikipedia, http://en.

wikipedia.org/wiki/Prime_number , accessed 2012).
Other methods for prime identification are rather complex for a savant to
reasonably learn, and many are unsuitable for a parallel algorithm. Also, based on
electroencephalography the mind is limited to about 40 images per second, so most
serial algorithms would take too long.
Many calculating tasks can be visualized as a massively parallel set of operations
followed by accumulation. For example, when multiplying two large numbers, the
multiplicand may be multiplied by each digit of the multiplier in parallel. Then all
that remains is to sum the partial products by adding them to an accumulation one at
a time to determine the product. Multiplying 678 by 345, for example, can be
accomplished by multiplying 678 by 3, while concurrently multiplying 678 by 4,
while concurrently multiplying 678 by 5; the product is simply the appropriate sum
of the partial products, taken with controlled toggles.
It is interesting to note that the above priority-comparator architecture might be
adapted to perform mental arithmetic. For instance, instead of priority, partial
products could be determined (in parallel) using codes from long-term memory to
activate controlled toggles in a logical register. Instead of a magnitude comparator,
an accumulator could be implemented. Thus the above system is capable, at least in
theory, of more than priority calculations.

Conclusions

Simulated qubits that operate subliminally perform the necessary digital arithmetic
by using routines in the form of codes taken from long-term memory. Such codes
flow without conscious exertion, and provide instructions that achieve calculations,
in this case integer additions to obtain a priority value, and integer subtractions to
accomplish priority comparison and selection.
The method described involves a sequence of operations in which fm toggles, if
all true, are made to cause to toggles to change states. Operations like this may sum
weights to obtain the priority for a given image. Similar operations using fm-to
operations may subtract priorities, to determine which is larger. Finally an image
multiplexer is assumed to gate the highest priority image into conscious short-term
memory.