Transposing of Matrices without creating a new one

[Sven Wirz]

Hello ladies and gentlemen,

I got a brief question: which is the best way to implement A^T * A * x, where A is a sparsematrix, A^T the transposed sparsematrix and x a given vector. I just doesn't seem efficient to create two Matrices to accomplish my goal.

sincerely

S. Wirz

[Vladimir Kostyukov]

Hello Sven,

I see what are looking for. Looks like "multiply matrix by its transpose" it's a popular thing. For now, it's not possible to do that w/o overhead. Even an upcoming in-place operations won't help here. The only way to do that is to create an additional tranposed matrix.

I've created a ticket regarding this request https://github.com/vkostyukov/la4j/issues/200 (bingo 200!). For now, I'm planing to handle this in 0.6.0 release.

Thank you for the request! Feel free to ask any further questions on la4j.

--
Thanks,

Vladimir Kostyukov