I am performing a Cholesky decomposition, like in this thread. Post #4 contains a minor example, that looks like as my code. The minor example of Ying is based on this example. As one can see, the matrix is read by the master node and then gets distributed (every element of the matrix). Then every element of the matrix is being gathered back to the master node.
However, the example is general. In my application, the matrix is symmetric, so I am wondering if I could do better, than just sending all the matrix. That is, to send only the lower part of tha matrix, perform the Cholesky decomposition and then receive only the lower part of the matrix. Is this possible?
Of course, this has to do on how pdpotrf() expects it's input. My idea is that if we could limit distribution only in the lower part of the matrix, then we could achieve better performace.
PS - a negative answer will be also flagged as the best reply, i.e. it is accepted.
