RE : [scilab-Users] Re: RE : [scilab-Users] inversion of a huge block diagonal matrix

[Collewet Guylaine]

I do not use this script to produce the matrix
I am sorry, it is not efficient at all ! 
I guess i should find a "vectoriel" way to do the same

GC

-------- Message d'origine--------
De: Collette yann [mailto:yann.collette at scilab.org]
Date: mar. 16/03/2010 16:35
À: users at lists.scilab.org
Objet : Re: [scilab-Users] Re: RE : [scilab-Users]  inversion of a huge block diagonal matrix
 
I performed a small computation:
a 5x5 matrix * 1e6 blocks * 8 octect = 200 Mo.
Plus some copies of this matrix (when you pass a parameter to a scilab 
function,  this parameter is copied).
So, we are near the memory limit of Scilab (1Go for the scilab-5 version).

Do you use this script to produce your matrix ? If yes, how much time 
does it takes on your PC ?

YC

Collette yann a écrit :
> Thanks for the script.
> Is it possible for you to perform the inversion of the matrix A while 
> generating this matrix ?
>
> YC
>
> Collewet Guylaine a écrit :
>>
>> yes,
>> please find attached the script which produces a block diagonal matrix
>>
>> GC
>>
>>
>> -------- Message d'origine--------
>> De: Collette yann [mailto:yann.collette at scilab.org]
>> Date: mar. 16/03/2010 15:59
>> À: users at lists.scilab.org
>> Objet : [scilab-Users] Re: RE : [scilab-Users] inversion of a huge 
>> block diagonal matrix
>>
>> This only other solution I see is to write a C-interface, to transfer
>> all the blocks from the diagonal and to perform the computation using C
>> functions.
>> Do you have a script which produce such a matrix ?
>>
>> YC
>>
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 3743 bytes
Desc: not available
URL: <https://lists.scilab.org/pipermail/users/attachments/20100316/b984e56e/attachment.bin>