I'm sending you this email for a question about your libraries.
I'm using your Matrix class as follow
the myArray variable is a one dimensional array that countain all the RGB of a given picture ( so if the picture countain 64 000 pixels, the array length will be 64 000, in reality the picture i have are even bigger . . . )
Matrix a = new Matric( myArray );
SinGularValueDecomposition svd = a.svd();
But every time i call svd(); with my Matrix, the JVM run out of memory.
I know that i have a huge amount of data, but i wanted to know if there is any way of not having this problem ?
I don't know if you have any other SVD implementation that could solve this problem ?
In my algorithm, i tried to minimise the use of memory, but no matter what i do i always use too much memory for the JVM.
Thank you if you can help me, and sorry for my poor english
Best regards,
Kostia Chardonnet.
- - - - - -
University Paris8