>
> People,
>
> I am trying to interface a C program fifoToA.c to Axiom (friCAS)
> by named pipes in Linux (Unix).
> These pipes are created by the system commands > mkfifo toA
> > mkfifo fromA
>
> fifoToA builds n strings "ABC0", "ABC1", "ABC2", ... in a loop,
> outputs each string to the pipe toA,
> inputs the respond string from the pipe fromA.
>
> fifoFromA.input is interpreted by Fricas.
> It inputs a string iStr from toA,
> converts each character in iStr to the lower case,
> outputs the result string to fromA.
>
> 1. C <-> C
> ----------
> First, I replace fifoFromA.input with the C program fifoFromA.c.
>
> fifoToA opens toA once and only for writing,
> opens fromA once and only for reading.
> fifoFromA.c opens toA once and only for reading,
> opens fromA once and only for writing.
>
> A particular point is that in both C programs the function `open'
> (`close') is envoced only once. And the C programs use read/write
> (somewhat of a lower level) instead of fgets/fputs
> (I do not know why fgets-fputs-fflush did not work as needed).
>
> The source is presented below.
>
> Compilation: gcc -O2 -o fifoToA fifoToA.c
> gcc -O2 -o fifoToA fifoToA.c
>
> First, run > ./fifoFromA
> on terminal-2,
> then, run > ./fifoToA
> on terminal-1.
>
> This C <-> C exchange performance (with no intermediate printing in
> the loop) at 133000 strings/sec on a 2 GHz machine
> (say, put n = 400000 in `main').
> It is slow (I need to find out, why).
It is about 10 times slower than my first guess, but within
reasonable margin. Google for 'pipe based context switching'
to learn about expected results.
> But so far, it looks admissible to start with and to experiment
> further with parsing a string and applying algebra at the Axiom end.
>
> 2. C <-> Axiom
> --------------
> Only I failed to replace ./fifoToA.c with a correct code in
> fifoFromA.input
>
> (first, run on terminal-2 fricas and )r fifoFromA.input
> Then, command > ./fifoToA
> on terminal-1
> ).
> Can you, please, fix the below code in fifoFromA.input ?
>
The fifoFromA.input that I posted works for me when using
trunk build using debian CLISP 2.48, after about 24 min it did 332000
iterations. Btw: I use C code with fix from the other message.
As I wrote, older sbcl (and this seem to include SBCL 1.0.40.0.debian)
has bug so the code will not work.
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/fricas-devel?hl=en.
