Re: [JDEV] c / c++?

["Corbett J. Klempay" <cklempay@chimera.acm.jhu.edu>]

Yeah, I was actually looking at the pages last night...it seems like a
strong contender....I grabbed the PS docs so I can print them out later...

The fact that it has asm optimized guts for several architectures is key,
too....all of the literature I've seen on this mentions the importance of
asm optimization of your core bigint routines, as they play such an
integral part...

------------------------------------------------------------------------------
Corbett J. Klempay			         Quote of the Week:
http://www.acm.jhu.edu/~cklempay       "A commune is where people join 
					together to share their lack of 
					wealth."
------------------------------------------------------------------------------

On Sun, 18 Apr 1999 qbradley@csc.UVic.CA wrote:

> 
> Have you considered the GNU MP library?  It is GPL, hence "free", and is
> supposed to be fast.  It is C (with optional bonus assembly for many
> platforms)
> 
> http://www.gnu.org/software/gmp/gmp.html
> 
> "The speed of GNU MP is believed to be faster than any other similar
> library"
> 
> Quetzalcoatl Bradley
> qbradley@csc.uvic.ca
> 
> On Mon, 19 Apr 1999, Corbett J. Klempay wrote:
> 
> > On Sun, 18 Apr 1999, Nicholas M. Kirsch wrote:
> > 
> > > 
> > > Traditionally C compilers have trouble linking with C++ libraries. This is
> > > because C++ allows for function overloading (ie. having foo (void) & foo
> > > (int) ) and therefore mangles the symbol table.
> > >
> > 
> > Yes, this is what I fear.  I am leaning towards doing whatever it takes to
> > keep it pure ANSI C...there are just too many headaches with mixing the
> > two.
> >  
> > > Not that there isn't a way around it, but have you looked into any ANSI C
> > > cryptography libraries.. I've played around with a couple that are good.
> > > Let me look them up and direct you to their sites.
> > > 
> > 
> > Sure thing.  Full crypto libraries are the nicest, but the most essential
> > thing I need is a solid large integer (and associated functions for
> > working in Z(n)) package.  I've looked into freelip (derived from
> > Lenstra's RSA129 effort)...but I like the fact that some of these other
> > ones have ASM code for a few different architectures at the root for
> > speed.  I have a superb book on elliptic curve crypto, for example, that
> > rolls its own bigints...but admits that they implementation is totally for
> > readability/understandability vs. efficiency...I'd like a good compromise
> > between the two.
> > 
> > CJK
> > 
> > 
> 
> Quetzalcoatl Bradley
> qbradley@csc.uvic.ca
>