Dear FriCAS developers,
This is on the named pipe string interface for Haskell -- C -- Axiom.
Something like a report and also a question.
I thank Waldek Hebisch who helped me with the current examples.
The actual example is on
http://botik.ru/pub/local/Mechveliani/axiQuest/report.zip
The pipes are made by > mkifo toA
> mkifo fromA
The first considered variant there is by using fgets/fputs
-- the one corrected by Waldek Hebisch.
It is in the root directory of the archive.
Let us call this variant
"fgets".
It works now for the part of C -- Axiom.
And fifoA.c does not apply fopen inside the loop.
This variant has the following remaining problem:
I failed to replace fifoA.c with the Haskell program.
It processes only the first string, and then gets to error.
Haskell has a rich IO library; named pipes are more subtle than
files, it is difficult to find out what is wrong.
Instead, the read-write variant (call it "read") has succeeded with
Haskell -- C (see below).
Here are the notes on the "fgets" variant for C -- Axiom.
1. Performance.
Consider the strings "ABC0", "ABC1", ... for n = 100000,
and fifoFromA.c applying lowercase to each string,
and fifoFromA.input applying lowerCase to each string,
and nothing else,
and no printing inside the loop
(and `count' is removed from fifoFromA.input).
Put the performance of C -- Axiom fgets to be 1.
Then: C -- C "read" has performance 22.
Haskell -- C "read" has performance 20.
C -- C fgets has performance 14.
C -- Axiom fgets has performance 1.
Is this ratio of 14 due to that the loop in Axiom is interpreted?
2. Dependence on the length
Taking a 2 times larger string does not effect the time in C -- C,
but it increases 3/2 the time in C -- Axiom.
3. Depenence on printing the counter
If fifoFromA.input prints out each 5000-th value of the counter,
this effects the time (if I have not mistaken).
The "read" variant
==================
Its source is in /byReadWrite
It is by using lower level functions open, read, write.
I tend to apply it because Haskell -- C works with it.
And now, it remains to make C -- Axiom to work with "read",
and then to replace C with Haskell.
And C -- Axiom works only "in half".
First run fifoFromA on terminal-2,
then run fifoA on terminal-1.
C -- C works as needed.
And C -- Axiom works as
n = 7
pair = ABC0
, any
pair = ABC1
,
pair = ABC2
, any
pair = ABC3
...
-- it performs correct each second iteration.
Can you, please, tell what I am missing?
Regards,
------
Sergei
[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.
