Re: TI-M: Re: TI-Math Digest V1 #22

[Prev][Next][Index][Thread]

Re: TI-M: Re: TI-Math Digest V1 #22



> I'm not sure on the specs specifically, but I would think binary floats
would
> be just as precise as BCD as long as approximately equal significant
digits
> are kept...

Is it IEEE that isn't fully precise?  Maybe I'm thinking of another format.
Anyways, it doesn't really matter =)

> A congruency list with modulus 2310
> consists of 485 (I think) probable-primes; I'm not sure how many of those
are
> composite, I might look into that just for curiosity's sake.

343    =)

> Anyway, to
> really make the prime list worth it, it would probably have to be
relatively
> large.

Not really worth it?  Say 2311 is a factor - you'd have to trial divide 485
probable primes by 2, 3, 5, 7, and 11.  That's gotta take a good bit of
time?

> Actually, that's a bit too much work for determining the number isn't
> divisible by 2, 3, 5, or 7, since the only numbers in a congruency list of
> mod 2310 are those that aren't divisible by 2, 3, 5, or 7...

don't forget 11.  I think that checking whether or not it's prime first is
worth it, since it's just a simple binary search - what is that, 12
iterations max?  When 70% of your probables ar primes, it's that much more
likely to save time. . .

> Give TI *some* credit, then ;)

Let's time it on a HP49G first =)

> >  Got any binary multiplication/division algorithms?
> Sure, these are fairly well-known algorithms:

good stuff =)

> I hope these are right, someone correct me if I screwed up somewhere ;)

They better be, since there doesn't appear to be too many other people here
who can =)

> Never mind, i guess it would have to in order to find the factors of 9 876
> 511 069 135 577...

hehe =P

> >  My only guess is that there's another algorithm, or some of that 2 megs
of
> >  Flash ROM is devoted to a prime list and wheeler is used afterwards. .
.
>
> I doubt the ROM would use a prime list to start out with and then wheeler,
> but you'd never know...

nah, TI probably had a list of the first 20,000 primes on AMS 1.00, then
decided to sacrifice mathematical speed for space on 2.03 - how else do they
free several hundred k's of mem?  =)

    -Scott
References: