Check return code (#329)

* Check rocBlas API result code in benchmark warmup calls

* Run clang-format on testing_gemm*hpp files

* Another formatting pas

Author: bensander

clients/include/testing_gemm.hpp

@@ -440,8 +440,8 @@ rocblas_status testing_gemm(Arguments argus)
 
         for(int i = 0; i < number_cold_calls; i++)
         {
-            rocblas_gemm<T>(
-                handle, transA, transB, M, N, K, &h_alpha, dA, lda, dB, ldb, &h_beta, dC, ldc);
+            CHECK_ROCBLAS_ERROR(rocblas_gemm<T>(
+                handle, transA, transB, M, N, K, &h_alpha, dA, lda, dB, ldb, &h_beta, dC, ldc));
         }
 
         gpu_time_used = get_time_us(); // in microseconds
@@ -626,7 +626,6 @@ rocblas_status range_testing_gemm(Arguments argus)
                 rocblas_error =
                     norm_check_general<T>('F', size, size, size, hC_gold.data(), hC.data());
             }
-
         } // end of if unit/norm check
 
         if(myfile.is_open())

clients/include/testing_gemm_strided_batched.hpp

@@ -281,24 +281,24 @@ rocblas_status testing_gemm_strided_batched(Arguments argus)
 
         for(int i = 0; i < number_cold_calls; i++)
         {
-            rocblas_gemm_strided_batched<T>(handle,
-                                            transA,
-                                            transB,
-                                            M,
-                                            N,
-                                            K,
-                                            &h_alpha,
-                                            dA,
-                                            lda,
-                                            stride_a,
-                                            dB,
-                                            ldb,
-                                            stride_b,
-                                            &h_beta,
-                                            dC,
-                                            ldc,
-                                            stride_c,
-                                            batch_count);
+            CHECK_ROCBLAS_ERROR(rocblas_gemm_strided_batched<T>(handle,
+                                                                transA,
+                                                                transB,
+                                                                M,
+                                                                N,
+                                                                K,
+                                                                &h_alpha,
+                                                                dA,
+                                                                lda,
+                                                                stride_a,
+                                                                dB,
+                                                                ldb,
+                                                                stride_b,
+                                                                &h_beta,
+                                                                dC,
+                                                                ldc,
+                                                                stride_c,
+                                                                batch_count));
         }
 
         gpu_time_used = get_time_us(); // in microseconds