What percentage of the time do you expect to be writing or reading from
fifos compared to computing?
The tactic of loading all of GMP and NTL into Maxima so you can just call
pieces has much lower overhead, but maybe that doesn't matter -- it has the
problem that loading foreign functions into Common Lisp is not completely
standardized. But maybe the pipes and sockets approach is not entirely
standard either.
RJF
> -----Original Message-----
> From: maxima-bounces at math.utexas.edu
> [mailto:maxima-bounces at math.utexas.edu] On Behalf Of Fabrizio Caruso
> Sent: Saturday, September 15, 2007 7:32 AM
> To: Andrej Vodopivec
> Cc: maxima at math.utexas.edu
> Subject: Re: [Maxima] writing into and reading from fifos
>
> Thanks!
>
> This works!
>
> Now I have a C++ server that performs very fast computations
> for Maxima or any other system that can write into named pipes.
>
> I would like to submit this as a contribuited package
> once I have written a decent Maxima interface.
> At the moment my main interest is having fast LLL computation.
> Maybe I'll also write a version that uses sockets.
>
> Fabrizio
>
>
>
> On Sat, 15 Sep 2007, Andrej Vodopivec wrote:
>
> > On 9/14/07, Fabrizio Caruso <caruso at dm.unipi.it> wrote:
> >> I would like to read and write strings into such files.
> >>
> >> I can open for writing a file as a stream by
> >> "f : openr(<path>)"
> >>
> >> I can read it by
> >> "readline(f)"
> >
> > To write to a file you need to open it with openw. To
> append stuff to
> > a file you need to open it with opena. To write stuff to a file use
> > printf.
> >
> >
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>