see.sig

Take a look at the Linpack benchmark, which is used in the Top 500 (dot
org). That's a system solution (solve for x in Ax=f where f is given).

Also take a look at SparseBench, which is a benchmark for operations on
sparse matrices.

The matrix-vector product with dense matrices is not overly interesting,
because it's not a dominant kernel in most scientific applications.

Getting an element is somewhat interesting, but usually people arrange
the matrix in such a way that all elements are already where they will
be needed in the future. I know of no benchmarks that test just this
operation.

V.
--
email: lastname at cs utk edu
homepage: www cs utk edu tilde lastname