From: <ctk...@o2...> - 2004-06-14 12:16:29
|
Hello, I am trying to implement function eigs(), as Matlab does it, using an Arp= ack library. Unfortunately I’ve stacked at the beginning and can't = move any step forward. Problem is as follows... Simple program using arpack written in C works OK. But almost the same c= ode used as a function in oct-file produces wrong results. For example ap= rack function asked for starting vector always returns [NaN NaN NaN ....] Of course, the same problem arise in C++ instead if C. I think of some memory problem. It looks like aprack functions are referr= ing to wrong place in memory when called from octave function. However, s= ome of them do it all right. I've debug Arpack some how, and notice that = i.e. this starting vector looks all right till it is normalized. =20 I wonder if: 1)anybody has faced similar problem? 2)anybody is interested in helping me? If so, I'll put some more details. Regards, Czarek Tkaczyk |