Fixed test suite flop formulas for ops with side.

Details:
- Fixed incorrect flop counts in test suite modules for hemm, symm, trmm,
  trmm3, and trsm.
- Comment updates in herk macro-kernels.

frame/3/herk/bl2_herk_l_ker_var2.c

@@ -250,7 +250,7 @@ void PASTEMAC(ch,varname)( \
 			diagoffc_ij = diagoffc - (doff_t)j*NR + (doff_t)i*MR; \
 \
 			/* If the diagonal intersects the current MR x NR submatrix, we
-			   compute in the temporary buffer and then add in the elements
+			   compute it the temporary buffer and then add in the elements
 			   on or below the diagonal.
 			   Otherwise, if the submatrix is strictly below the diagonal,
 			   we compute and store as we normally would.

frame/3/herk/bl2_herk_u_ker_var2.c

@@ -250,7 +250,7 @@ void PASTEMAC(ch,varname)( \
 			diagoffc_ij = diagoffc - (doff_t)j*NR + (doff_t)i*MR; \
 \
 			/* If the diagonal intersects the current MR x NR submatrix, we
-			   compute in the temporary buffer and then add in the elements
+			   compute it the temporary buffer and then add in the elements
 			   on or below the diagonal.
 			   Otherwise, if the submatrix is strictly above the diagonal,
 			   we compute and store as we normally would.

testsuite/src/test_hemm.c

@@ -222,10 +222,7 @@ void libblis_test_hemm_experiment( test_params_t* params,
 	}
 
 	// Estimate the performance of the best experiment repeat.
-	if ( bl2_is_left( side ) )
-		*perf = ( 1.0 * m * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
-	else // if ( bl2_is_right( side ) )
-		*perf = ( 1.0 * m * n * n ) / time_min / FLOPS_PER_UNIT_PERF;
+	*perf = ( 2.0 * mn_side * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
 	if ( bl2_obj_is_complex( c ) ) *perf *= 4.0;
 
 	// Perform checks.

testsuite/src/test_symm.c

@@ -222,10 +222,7 @@ void libblis_test_symm_experiment( test_params_t* params,
 	}
 
 	// Estimate the performance of the best experiment repeat.
-	if ( bl2_is_left( side ) )
-		*perf = ( 1.0 * m * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
-	else // if ( bl2_is_right( side ) )
-		*perf = ( 1.0 * m * n * n ) / time_min / FLOPS_PER_UNIT_PERF;
+	*perf = ( 2.0 * mn_side * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
 	if ( bl2_obj_is_complex( c ) ) *perf *= 4.0;
 
 	// Perform checks.

testsuite/src/test_trmm.c

@@ -212,10 +212,7 @@ void libblis_test_trmm_experiment( test_params_t* params,
 	}
 
 	// Estimate the performance of the best experiment repeat.
-	if ( bl2_is_left( side ) )
-		*perf = ( 1.0 * m * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
-	else // if ( bl2_is_right( side ) )
-		*perf = ( 1.0 * m * n * n ) / time_min / FLOPS_PER_UNIT_PERF;
+	*perf = ( 1.0 * mn_side * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
 	if ( bl2_obj_is_complex( b ) ) *perf *= 4.0;
 
 	// Perform checks.

testsuite/src/test_trmm3.c

@@ -223,10 +223,7 @@ void libblis_test_trmm3_experiment( test_params_t* params,
 	}
 
 	// Estimate the performance of the best experiment repeat.
-	if ( bl2_is_left( side ) )
-		*perf = ( 1.0 * m * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
-	else // if ( bl2_is_right( side ) )
-		*perf = ( 1.0 * m * n * n ) / time_min / FLOPS_PER_UNIT_PERF;
+	*perf = ( 1.0 * mn_side * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
 	if ( bl2_obj_is_complex( c ) ) *perf *= 4.0;
 
 	// Perform checks.

testsuite/src/test_trsm.c

@@ -213,10 +213,7 @@ void libblis_test_trsm_experiment( test_params_t* params,
 	}
 
 	// Estimate the performance of the best experiment repeat.
-	if ( bl2_is_left( side ) )
-		*perf = ( 1.0 * m * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
-	else // if ( bl2_is_right( side ) )
-		*perf = ( 1.0 * m * n * n ) / time_min / FLOPS_PER_UNIT_PERF;
+	*perf = ( 1.0 * mn_side * m * n ) / time_min / FLOPS_PER_UNIT_PERF;
 	if ( bl2_obj_is_complex( b ) ) *perf *= 4.0;
 
 	// Perform checks.