Thank you Adrien,
I tried your approach before because it works with optim. It fails. The reason
is that "x" is a variable of fminsearch which interferes with "my" x.
The consequence is that one has to choose variables that are not defined in
fminsearch. This is quite disappointing for an elaborat
imagine the following :
for i = 1 : 10
for j = 1 : 8
lanch_my_function_on_point_j
end
end
would like to do :
for i = 1 : 10
proc1 = my_function_on_point_1
proc2 = my_function_on_point_2
proc3 = my_function_on_point_8
end
Paul
NB : the
There is no perfect answer: It depends the problem you are trying to solve.
Sylvestre
On 02/12/2012 15:38, paul.carr...@free.fr wrote:
> Hi Sylvestre,
>
> Let me asking the following : in order to use parallelization in a multicore
> environnent (serveur or station), what is the best way ?
>
>
Hi Sylvestre,
Let me asking the following : in order to use parallelization in a multicore
environnent (serveur or station), what is the best way ?
NB :
- As I ever told in the current list, I use as often as possible the
vectorization in my current develoment, but the same elementary scheme h
On 02/12/2012 11:59, Martin Helm wrote:
> Am 02.12.2012 11:48, schrieb paul.carr...@free.fr:
>> I don't know how to say ... I'm using the 5.4.0 binary release and I've such
>> kind of message
>>
>> see
>> http://mailinglists.scilab.org/Scilab-users-parallilization-error-on-opensuse-td4025260
thanks for your suggestion ; unfortunately nothing changes in my opensuse OS ...
I had a look in a Suse package but the latest Scilabk is not available ...
it's a pitty
Paul
#"""
On 02/12/2012 09:33, Jens Simon Strom wrote:
Hello Scilab experts,
This script functions:
function SAQ=!f(p), SAQ=sum( ( p(1)+p(2)*xi-yi ).^2 ), endfunction
xi=1:1000 ;
yi=2*xi+3;
popt=fminsearch(!f,[0;0])
If I replace yi by y everywhere the script functions too as expected.
If I replace xi b
>From the bug report you quoted earlier I take that at least on openSUSE
it can be what I suspected in the first place. You miss a library which
is needed by libparallel.
Please post from your openSUSE system the output from
ldd /home/paul/scilab-5.4.0/lib/scilab/libsciparallel.so
and
ls -l /hom
Am 02.12.2012 11:48, schrieb paul.carr...@free.fr:
> I don't know how to say ... I'm using the 5.4.0 binary release and I've such
> kind of message
>
> see
> http://mailinglists.scilab.org/Scilab-users-parallilization-error-on-opensuse-td4025260.html
>
> Paul
>
That looks very suspicious, th
I don't know how to say ... I'm using the 5.4.0 binary release and I've such
kind of message
see
http://mailinglists.scilab.org/Scilab-users-parallilization-error-on-opensuse-td4025260.html
Paul
- Mail original -
De: "Martin Helm"
À: users@lists.scilab.org
Envoyé: Dimanche 2 Déce
I am a bit surprised that you have a problem on openSUSE, it works just
fine for me on openSUSE 12.2 64 bit, on Centos I cannot comment.
___
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
Unfortunately after deeper research on the net, it seems necessary to build
Scilab when such error message appears, in order to use parallel_run
function ...
http://permalink.gmane.org/gmane.comp.mathematics.scilab.user/6379
http://bugzilla.scilab.org/show_bug.cgi?id=9381
http://mailinglists.scila
Am 02.12.2012 09:37, schrieb Paul Carrico:
>
> Dear,
>
> I installed Scilab binary files under 2 different linux OS (Opensuse
> and Centos) ; in both cases the libsciparallel.so library is missing …
>
> is there any workaround in order to use parallel_run function ?
>
> (I haven’t seen any rpm file
Dear,
I installed Scilab binary files under 2 different linux OS (Opensuse and
Centos) ; in both cases the libsciparallel.so library is missing .
is there any workaround in order to use parallel_run function ?
(I haven't seen any rpm file of something like this .)
A subsidiary question
Hello Scilab experts,
This script functions:
function SAQ=!f(p), SAQ=sum( ( p(1)+p(2)*xi-yi ).^2 ), endfunction
xi=1:1000 ;
yi=2*xi+3;
popt=fminsearch(!f,[0;0])
If I replace yi by y everywhere the script functions too as expected.
If I replace xi by x everywhere an error occurs: "Inconsistent s
15 matches
Mail list logo