I wrote:
>
> Serge D. Mechveliani wrote:
> >
> > People,
> >
> > can you, please, answer the following beginner questions on compilation
> > and performance
> > (for FriCAS-1.1.5 - GNU CLISP 2.48 - Debian Linux).
>
> I tried 5 versions of your code on different Lisp versions:
>
> Clisp ecl sbcl
> ct1(4*10^5) 0.66 0.18 0.08
> ct2(4*10^5) 0.72 0.18 0.07
> ct3(4*10^5) 5.59 1.08 2.04
> ct4(4*10^5) 0.64 0.13 0.05
> ct5(4*10^5, all) 0.64 0.14 0.05
>
I added another one using '#' for string length:
clisp ecl sbcl
ct6(4*10^5) 0.06 0.02 0.01
ct6(4*10^7) 5.96 1.28 0.49
As you can see this one is about 10 time faster (no wonder,
length is stored together with characters and '#' only has
to read it from meomory). For sbcl '#' takes only 12% of
execution time (the rest is arithmetic on 'i' and
loop overhead).
BTW: All your version use bignums for arithmetic and loop
control. You can make it faster restricting arguments
to SingleInteger. In interpreter this gives little gain
(of order 30%), but the following Spad version (a little
variation of Ralf code) is much faster:
)abbrev package COUNT Count
Count(): with
ct: SingleInteger -> SingleInteger
== add
ct(n: SingleInteger) : SingleInteger ==
str: String := "abc"
i := n
while 0 < i repeat
i := i - (#str pretend SingleInteger)
return i
with sbcl ct(4*10^7) is 0.02 sec while ct(4*10^8) take 0.14 sec.
So another factor of 30.
--
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.
