Re: Calaulation problem

You wrote:

"completely factor each number and store each
member of the pair AS an array of prime numbers"

Which I pointed out in many circumstances this is either impossible or would
take much too long to evaluate to be a viable system and simply would not
work for non-integers, so where is the problem of relevance? I think that my
comment was relevant to what you said.

What it seems is being suggested is that some numbers be stored as a n/d
pair, others as an array of primes. So instead of an approximation with
accuracy determined by the choice of data type, the computer would take a
guess and store each number in the way that seemed best at the time. This
would be better how?

Dave O.

"Bob O`Bob" <filterbob@xxxxxxxxxxxxxxx> wrote in message
news:eTQJqkzKGHA.1088@xxxxxxxxxxxxxxxxxxxxxxx
Dave wrote:

Also that will only work for integers, you will never find any primes
which when multiplied together result in a non-integer.


...and what thread have YOU been reading? Certainly not /this/ one.


.

Relevant Pages

  • Re: hash table idea good or no good?
    ... It unsigned long is 32 bits and the primes I listed are much smaller ... You want to use strings as the key. ... You want to store strings, but you use the entire string as one big ... least common multiple scheme used above. ...
    (comp.programming)
  • Re: greatest multiple algorithm
    ... array stores found primes ... modulo with found primes in array to find new primes ... consumes large amounts of memory ... That doesn't sound like a sieve, ...
    (comp.lang.asm.x86)
  • Re: [Algorithm] Sum of Primes < 1000000
    ... Note also skipping the inner loop for composite values of i. ... The number of primes is O) from what I've seen of primes, which is Osince the log of the square root is simply half the log and the half disappears into the constant factor. ... The inner loop does one imul, three iadds (one's actually the subtract in the loop test and one computes the array index), one icmp, and one boolean assignment. ...
    (comp.lang.java.programmer)
  • Re: Vernons Prime Sieve
    ... // your primesarray is not big enough. ... // the primes that fit in 16 bits can be used to find ... > (i.e. it takes about twice as long to do test twice as ... would like hexadecimal output, then %X is the way to go. ...
    (sci.math)
  • Re: More on triangle numbers and primes!
    ... > want to extract or the limit of the array or file you have for storage. ... > spots by identifying the location as a prime or if an ... > for listing out the primes. ... Hmm...Do you have a pattern that runs in accelerating numerical value steps ...
    (sci.math)