Add tests for bml_add_norm

Closes: lanl#531
Signed-off-by: Nicolas Bock <[email protected]>

tests/C-tests/add_matrix_typed.c

@@ -20,11 +20,13 @@ int TYPED_FUNC(
     const int M)
 {
     bml_matrix_t *A = NULL;
-    bml_matrix_t *B = NULL;
+    bml_matrix_t *B_1 = NULL;
+    bml_matrix_t *B_2 = NULL;
     bml_matrix_t *C = NULL;
 
     REAL_T *A_dense = NULL;
-    REAL_T *B_dense = NULL;
+    REAL_T *B_1_dense = NULL;
+    REAL_T *B_2_dense = NULL;
     REAL_T *C_dense = NULL;
 
     bml_distribution_mode_t distrib_mode = sequential;
@@ -39,23 +41,29 @@ int TYPED_FUNC(
     double alpha = 1.2;
     double beta = 0.8;
     double threshold = 0.0;
+    double trace_norm = 0.0;
 
     int max_row = MIN(N, PRINT_THRESHOLD);
     int max_col = MIN(N, PRINT_THRESHOLD);
 
     LOG_DEBUG("rel. tolerance = %e\n", REL_TOL);
 
     A = bml_random_matrix(matrix_type, matrix_precision, N, M, distrib_mode);
-    B = bml_copy_new(A);
+    B_1 = bml_copy_new(A);
+    B_2 = bml_copy_new(A);
     C = bml_random_matrix(matrix_type, matrix_precision, N, M, distrib_mode);
 
     LOG_DEBUG("bml_add...\n");
-    bml_add(B, C, alpha, beta, threshold);
+    bml_add(B_1, C, alpha, beta, threshold);
+
+    LOG_DEBUG("bml_add_norm...\n");
+    trace_norm = bml_add_norm(B_2, C, alpha, beta, threshold);
 
     LOG_INFO("alpha = %f \n ", alpha);
     LOG_INFO("beta = %f \n ", beta);
     A_dense = bml_export_to_dense(A, dense_row_major);
-    B_dense = bml_export_to_dense(B, dense_row_major);
+    B_1_dense = bml_export_to_dense(B_1, dense_row_major);
+    B_2_dense = bml_export_to_dense(B_2, dense_row_major);
     C_dense = bml_export_to_dense(C, dense_row_major);
 
     if (bml_getMyRank() == 0)
@@ -67,29 +75,43 @@ int TYPED_FUNC(
         bml_print_dense_matrix(N, matrix_precision, dense_row_major, C_dense,
                                0, max_row, 0, max_col);
         LOG_INFO("B = alpha A + beta C \n");
-        bml_print_dense_matrix(N, matrix_precision, dense_row_major, B_dense,
+        bml_print_dense_matrix(N, matrix_precision, dense_row_major, B_1_dense,
                                0, max_row, 0, max_col);
 
+        double expected_trace_norm = 0.0;
         for (int i = 0; i < N * N; i++)
         {
             double expected = alpha * A_dense[i] + beta * C_dense[i];
-            double rel_diff_val = (expected - B_dense[i]) / expected;
+            double rel_diff_val = (expected - B_1_dense[i]) / expected;
             double rel_diff = fabs(rel_diff_val);
             if (rel_diff > REL_TOL)
             {
                 LOG_ERROR
                     ("matrices are not identical; expected[%d] = %e, B[%d] = %e\n",
-                     i, expected, i, B_dense[i]);
+                     i, expected, i, B_1_dense[i]);
                 return -1;
             }
+            expected_trace_norm += expected * expected;
         }
+        expected_trace_norm = sqrt(expected_trace_norm);
+        double rel_diff_val = (expected_trace_norm - trace_norm) / expected_trace_norm;
+        double rel_diff = fabs(rel_diff_val);
+        if (rel_diff > REL_TOL)
+          {
+            LOG_ERROR
+              ("trace norm is incorrect; expected = %e but found = %e\n",
+               expected_trace_norm, trace_norm);
+            return -1;
+          }
         bml_free_memory(A_dense);
-        bml_free_memory(B_dense);
+        bml_free_memory(B_1_dense);
+        bml_free_memory(B_2_dense);
         bml_free_memory(C_dense);
     }
 
     bml_deallocate(&A);
-    bml_deallocate(&B);
+    bml_deallocate(&B_1);
+    bml_deallocate(&B_2);
     bml_deallocate(&C);
 
     LOG_INFO("add matrix test passed\n");