On Mon, Jul 04, 2022 at 12:03:11PM +0800, Qian Yun wrote:
> >>For bytecode compiler, ")time on" reports time that is twice of
> >>the actual time.
> >
> >We base our time report on Lisp 'get-internal-run-time'.
> >Could you try adding something like
> >
> >)lisp (get-internal-run-time)
> >
> >before and after. And also
> >
> >)lisp INTERNAL-TIME-UNITS-PER-SECOND
> >
> >Our time report should be close to difference of values
> >reported by 'get-internal-run-time' and converted to
> >normal units based on INTERNAL-TIME-UNITS-PER-SECOND.
> >
>
> For ECL bytecode, I observed that it runs in parallel somehow:
>
> One thread takes 100% CPU and rest threads take a few percent
> CPU (total is around 100%), so together it takes 200% CPU.
>
> I wonder if it is the same on your machine.
> My ECL version is 21.2.1.
I did a simple test, in file 'll5' I have had:
(load "fricas.lisp")
integrate((F^((-a*x+1)^(1/2)/(a*x+1)^(1/2)))^n/(-a^2*x^2+1),x)
)quit
Then I run:
time /mnt/lv0/fricas/usr-16.1.2/bin/ecl < ll5 > rapp12 2>&1
and similar for Clozure CL 1.11 and sbcl-2.2.6 with fasteval.
Results are:
ECL 16.1.2
real 0m16.018s
user 0m23.464s
sys 0m0.124s
sbcl-2.2.6 with sb-fasteval
real 0m22.867s
user 0m22.348s
sys 0m0.480s
Clozure CL 1.11
real 0m25.276s
user 0m25.080s
sys 0m0.120s
So, in real time ECL was fastest, but it shows slightly bigger
CPU time than sbcl-2.2.6. At least for use with FriCAS ECL-16.1.2
seem to be few percent faster than newer versions.
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/fricas-devel/20220706235111.GA12023%40fricas.math.uni.wroc.pl.
