1C = P^T * B: 2Mat Object: 2 MPI processes 3 type: mpiaij 4 rows=2, cols=5 5 total: nonzeros=10, allocated nonzeros=10 6 total number of mallocs used during MatSetValues calls=0 7 not using I-node (on process 0) routines 8 9C = P^T * B after MatProductClear(): 10Mat Object: 2 MPI processes 11 type: mpiaij 12 rows=2, cols=5 13 total: nonzeros=10, allocated nonzeros=10 14 total number of mallocs used during MatSetValues calls=0 15 not using I-node (on process 0) routines 16 17C = P^T * A * P: 18Mat Object: 2 MPI processes 19 type: mpiaij 20 rows=2, cols=2 21 total: nonzeros=2, allocated nonzeros=2 22 total number of mallocs used during MatSetValues calls=0 23 using nonscalable MatPtAP() implementation 24 not using I-node (on process 0) routines 25 26C = P^T * A * P after MatProductClear(): 27Mat Object: 2 MPI processes 28 type: mpiaij 29 rows=2, cols=2 30 total: nonzeros=2, allocated nonzeros=2 31 total number of mallocs used during MatSetValues calls=0 32 not using I-node (on process 0) routines 33 34Cdup: 35Mat Object: 2 MPI processes 36 type: mpiaij 37 rows=2, cols=2 38 total: nonzeros=2, allocated nonzeros=2 39 total number of mallocs used during MatSetValues calls=0 40 not using I-node (on process 0) routines 41