- General
-
Bugs
0
Fixed
reshape sparse matrix
Hello,
it seems that there is a problem when trying to reshape mp sparse matrices. For example, the code
toto=mp(rand(16,16));
reshape(toto,4,4^3)
does work (not a sparse matrix). But
toto=mp(sprand(16,16,0.3));
reshape(toto,4,4^3)
crashes Matlab (has to be closed and restarted completely). Would it be possible to correct this, as reshaping is one of the basic tools ?
Thanks,
Adam
it seems that there is a problem when trying to reshape mp sparse matrices. For example, the code
toto=mp(rand(16,16));
reshape(toto,4,4^3)
does work (not a sparse matrix). But
toto=mp(sprand(16,16,0.3));
reshape(toto,4,4^3)
crashes Matlab (has to be closed and restarted completely). Would it be possible to correct this, as reshaping is one of the basic tools ?
Thanks,
Adam
Customer support service by UserEcho
Thank you for the report.
Will add reshape & cat for sparse matrices shortly.
Apologies for delay with 'eigs'.
A month ago I found new approaches to substantially improve the performance of computations with dense matrices.
Have been implementing it since then - just finished the last step few days ago.
Toolbox now is much faster, even in quadruple mode.
Now continue working on sparse matrices and 'eigs'.
Now toolbox includes sparse reshape function: http://goo.gl/btsBt
Also it can be used on old Linux systems (starting from Red Hat 4.0).
could you also give me access to the OSX version ? Does not seem to work with the link you gave m above.
Thanks,
Adam
OSX is being compiled now. What exactly doesn't work? reshape?
Undefined function 'mpimpl' for input arguments of type 'double'.
Error in mp.Digits (line 118)
[varargout{1:nargout}] = mpimpl(1,varargin{:});
But maybe that's because this version is not for OSX ?
A.
OSX with reshape will be ready in few hours (compilation with all optimizations takes long time)
Thanks,
A.
Thanks
A.