Hello,<div>  I have a sparse matrix A. Its size is 494 x 494. It has 1666 nonzeros.</div><div>  I call MatMult(A, x, y).  From the profiling result (got by -log_summary), I find MatMult&#39;s flops is 2838.</div><div>  I think the flops should be nnz x 2 = 1666 x 2 = 3332.  I don&#39;t know how PETSc got this number (2838).</div>
<div>  I also observed the phenomena with other matrices. It seems PETSc always give smaller flops.  Why?</div><div><div>  </div><div> Thank you!</div>-- Junchao Zhang<br>
</div>