nicer output of the execution (adding GPU workload values)

bioem.cpp

@@ -615,7 +615,7 @@ int bioem::run()
 		    (((double) param.param_device.NumberPixels - (double) param.param_device.maxDisplaceCenter / 2.) * ((double) param.param_device.NumberPixels - (double) param.param_device.maxDisplaceCenter / 2.) * 2. + 8.) * (double) sizeof(myfloat_t) / compTime;
 		  const double nGBs2 = (double) RefMap.ntotRefMap * ((double) param.param_device.NumberPixels * (double) param.param_device.NumberPixels + 8.) * (double) sizeof(myfloat_t) / compTime;
 
-		  printf("\t\tTime Comparison %d %d: %f sec (%f GFlops, %f GB/s (cached), %f GB/s) (rank %d)\n", iOrient, iConv, compTime, nFlops / 1000000000., nGBs / 1000000000., nGBs2 / 1000000000., mpi_rank);
+		  printf("\t\tTime Comparison %d %d: %f sec (%f GFlops, %f GB/s (cached), %f GB/s, with GPU workload %d%%) (rank %d)\n", iOrient, iConv, compTime, nFlops / 1000000000., nGBs / 1000000000., nGBs2 / 1000000000., workload, mpi_rank);
 		}
 	      if (Autotuning == 1 && !stopTuning && STABLE_ITERATION(iConv))
 		{
@@ -688,7 +688,14 @@ int bioem::run()
 		      x = (c-b > b-a) ? (int)(b+(c-b)/2) : (int)(a+(b-a+1)/2);
 		    }
 
-		  if ( (c-b == limit) && ( b-a == limit) ) stopTuning=true;
+		  if ((c - b == limit) && (b - a == limit))
+		    {
+		      stopTuning=true;
+		      if (DebugOutput >= 2)
+			{
+			  printf("\t\tOptimal GPU workload %d%% (rank %d)\n", workload, mpi_rank);
+			}
+		    }
 
 		  workload=x;
 		  rebalance(x);

bioem_cuda.cu

@@ -684,7 +684,7 @@ void bioem_cuda::rebalance(int workload)
 
 	if (DebugOutput >= 2)
 	{
-	  printf("\t\tSetting GPU workload to %d%%\n", workload);
+	  printf("\t\tSetting GPU workload to %d%% (rank %d)\n", workload, mpi_rank);
 	}
 
 	GPUWorkload = workload;