Augustin Yiptong comments

Results 5 comments of


                                            Augustin Yiptong

sparse matrix * sparse vector is slow when the vector is relatively dense

Hi, thanks for your response. My application is pricing/managing/reporting reinsurance contracts. We essentially have a simulation with a very large number of events which can potentially take losses. My matrix...

sparse matrix * sparse vector is slow when the vector is relatively dense

Chewed a bit more on this and here is a better solution for my specific use case. My `A` matrix stays the same and i have to perform many multiplications...

sparse matrix * sparse vector is slow when the vector is relatively dense

Hi @Sacha0 I think it should be: `sgwi, sgwimax = 0, length(aux.sgw)` ?

sparse matrix * sparse vector is slow when the vector is relatively dense

Hi @Sacha0 `iszero(sgwv) && (resnnz += 1)` is not foolproof. Consider: ``` A = sparse(ones(1,3)) x = sparsevec([1.0, -1.0, 1.0]) aux = spmatspvec_init(A, Float64) spmatspvec(A,x,aux) ``` gives ``` Sparse vector...

sparse matrix * sparse vector is slow when the vector is relatively dense

No official fix.