Fwd: Maxima wishlist



--Alt-Boundary-26452.2369747
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Content-description: Mail message body

Am 27 Nov 2005 um 14:06 hat Richard Fateman geschrieben:

> There are lisp systems that use GMP for bignum arithmetic.
> In which case all these are available, I think.
> 
> There are lisp systems that do not use GMP for bignum arithmetic.
> They are probably faster for small number computations (64 bits or less,
> and maybe much faster for 28 bits or less), but are not really
> competitive with GMP for dealing with big numbers.
> 
> So it does not make sense to build these on top of Maxima, I
> suspect.
> Either they are there underneath, or are too slow to be
> seriously interesting.
> 
> Any thoughts from others?
> 

GCL-Maxima has an integer function gcd available at lisp level, which is an 
implementation of GMP-gcd. It is absolutely fast when using big numbers. This should 
be the standard for every integer function in Maxima. 

The code for this can be read in the GCL source files or in GMP directly. Unfortunately I 
am not very used to read C-code. Can perhaps someone copy me the original papers 
from Sorenson, Weber and Jebelean et al. with algorithms in pseudocode? I was not 
very lucky in searching for this articles. 

Volker van Nek

--Alt-Boundary-26452.2369747
Content-type: text/html; charset=US-ASCII
Content-transfer-encoding: 7BIT
Content-description: Mail message body






Am 27 Nov 2005 um 14:06 hat Richard Fateman geschrieben:

> There are lisp systems that use GMP for bignum arithmetic.
> In which case all these are available, I think.
>
> There are lisp systems that do not use GMP for bignum arithmetic.
> They are probably faster for small number computations (64 bits or less,
> and maybe much faster for 28 bits or less), but are not really
> competitive with GMP for dealing with big numbers.
>
> So it does not make sense to build these on top of Maxima, I
> suspect.
> Either they are there underneath, or are too slow to be
> seriously interesting.
>
> Any thoughts from others?
>

GCL-Maxima has an integer function gcd available at lisp level, which is an implementation of GMP-gcd. It is absolutely fast when using big numbers. This should be the standard for every integer function in Maxima.

The code for this can be read in the GCL source files or in GMP directly. Unfortunately I am not very used to read C-code. Can perhaps someone copy me the original papers from Sorenson, Weber and Jebelean et al. with algorithms in pseudocode? I was not very lucky in searching for this articles.

Volker van Nek
--Alt-Boundary-26452.2369747--