Has anyone made that BigInteger class

I was thinking of using D for one of my projects--the problem is that BigInteger is a vital component for it... has anyone implemented BigInt in D itself and could point me to a link? I'd be much obliged--I'm not sure implementing BigInt myself is my cuppa joe^H^H^Htea.

> has anyone implemented BigInt in D itself and could point me to a link? "Arcane Jill" posted a few days ago saying (s)he was writing one and that (s)he expected to finish it in a week or two... KTC -- Experience is a good school but the fees are high. - Heinrich Heine

In article <c8u62s$1fh9$1@digitaldaemon.com>, hellcatv@hotmail.com says... > >I was thinking of using D for one of my projects--the problem is that BigInteger is a vital component for it... > >has anyone implemented BigInt in D itself and could point me to a link? I'd be much obliged--I'm not sure implementing BigInt myself is my cuppa joe^H^H^Htea. Yes, me. I'm releasing it next week. It's destined to be called "etc.bigint". Right now it's still in development. Everything works, but I've got a couple more high level functions to add, and a little bit of optimizing and profiling to do. Can you wait a week? What do you need it FOR, by the way? I'll be doing a crypto library next, so if you want it for cryptography you could just wait even longer (or collaborate). Arcane Jill

May 25, 2004

Re: Has anyone made that BigInteger class

Posted by Daniel Horn
in reply to Arcane Jill

Permalink

Daniel Horn

Posted in reply to Arcane Jill

Permalink

Hi-
Actually I'm on a bit of a time budget here... so if you could send it t o me sooner I'd be much obliged (the hotmail addy is fine).
I only need +,-,*,/,%

I'd also help you beta test it.

I'm writing a ray tracer that is very precise (keeps track of intersections at rational locations)
It seems silly, but I can give you more motivational details when it's done. basically it's verifying the potential of using a ray tracer for more than raytracing ;-)
Currently I'm checkin it into dsource.org (project deliria)

--Daniel

Arcane Jill wrote:
> In article <c8u62s$1fh9$1@digitaldaemon.com>, hellcatv@hotmail.com says...
> 
>>I was thinking of using D for one of my projects--the problem is that BigInteger
>>is a vital component for it...
>>
>>has anyone implemented BigInt in D itself and could point me to a link? I'd be
>>much obliged--I'm not sure implementing BigInt myself is my cuppa joe^H^H^Htea.
> 
> 
> Yes, me.
> 
> I'm releasing it next week. It's destined to be called "etc.bigint".
> 
> Right now it's still in development. Everything works, but I've got a couple
> more high level functions to add, and a little bit of optimizing and profiling
> to do. Can you wait a week?
> 
> What do you need it FOR, by the way? I'll be doing a crypto library next, so if
> you want it for cryptography you could just wait even longer (or collaborate).
> 
> Arcane Jill
> 
> 
> 
>

Arcane Jill wrote: > Yes, me. > > I'm releasing it next week. It's destined to be called "etc.bigint". > > Right now it's still in development. Everything works, but I've got a couple > more high level functions to add, and a little bit of optimizing and profiling > to do. Can you wait a week? > > What do you need it FOR, by the way? I'll be doing a crypto library next, so if > you want it for cryptography you could just wait even longer (or collaborate). > > Arcane Jill What kind of algorithms are you planing to implement? Stephan

In article <c93j93$ukc$1@digitaldaemon.com>, Stephan Wienczny says... >What kind of algorithms are you planing to implement? > >Stephan "planning to" impies I haven't done it yet. Should be past tense, in fact. Low level routines are assember optimized. At the higher level, multiplication uses special case algorithms for squaring and multiplication by powers of two, and the Karatsuba-Ofman divide-and-conquer algorithm for very large numbers. Long division uses a radix 2^32 (rather than binary) method. Modulo exponentiation uses Mongomery reduction for odd moduli, and Barrett reduction for even moduli. Factorial and radix convertions do as much work as possible in uints. Modulo inversion uses the fast binary algorithm if the modulus is prime. Primality testing uses a fast lookup table for small numbers, and the Legendre test for larger numbers. (I may also give it a Rabin-Miller test but I haven't done that yet). Square root uses the shift and subtract method (not successive approximation). I may have forgotten to list a few. Other speedups are possible, but I'd say that's pretty ok for a first release. Arcane Jill

In article <c904uv$1skt$1@digitaldaemon.com>, Daniel Horn says... > >Hi- >Actually I'm on a bit of a time budget here... so if you could send it t >o me sooner I'd be much obliged (the hotmail addy is fine). >I only need +,-,*,/,% Hi, Not really, I'm afraid. All there is left for me to do now is wrapping it up, adding documentation, licensing (BSD-style) and so on. To be honest, it would take almost as long to wrap it up for you as for everybody else. Today is Wednesday. The package will be ready sometime over the weekend. If that's really not soon enough for your time budget then you can always use the gmp library (Gnu Multi-Precision) in C. Sorry, but, I'm an artist. Things will be ready when they are ready. That's all I can say. Arcane Jill

Arcane Jill wrote: > In article <c93j93$ukc$1@digitaldaemon.com>, Stephan Wienczny says... > > > >>What kind of algorithms are you planing to implement? >> >>Stephan > > > "planning to" impies I haven't done it yet. Should be past tense, in fact. > I wanted to know which _crypto_ alogrithm you are going to implement!?! RSA, ECC, Rijndael? Stephan

> Today is Wednesday. The package will be ready sometime over the weekend. If that's really not soon enough for your time budget then you can always use the gmp library (Gnu Multi-Precision) in C. If you do want to use GMP I've already made the D interface file and object wrappers. Get them at code: http://home.comcast.net/~benhinkle/gmp-d/gmp-d.zip documentation: http://home.comcast.net/~benhinkle/gmp-d/doc/

Forums