4 files changed, 24 insertions, 8 deletions
diff --git a/Microbenchmarks/STREAM/dpu/add.c b/Microbenchmarks/STREAM/dpu/add.c
index faf6504..12854ee 100644
--- a/Microbenchmarks/STREAM/dpu/add.c
+++ b/Microbenchmarks/STREAM/dpu/add.c
@@ -18,12 +18,16 @@ __host dpu_results_t DPU_RESULTS[NR_TASKLETS];
 
 // Add
 static void add_dpu(T *bufferC, T *bufferA, T *bufferB) {
-
+#if UNROLL
     #pragma unroll
     for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
         bufferC[i] = bufferA[i] + bufferB[i];
     }
-
+#else
+    for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
+        bufferC[i] = bufferA[i] + bufferB[i];
+    }
+#endif
 }
 
 // Barrier
diff --git a/Microbenchmarks/STREAM/dpu/copyw.c b/Microbenchmarks/STREAM/dpu/copyw.c
index eff7a3b..66dd319 100644
--- a/Microbenchmarks/STREAM/dpu/copyw.c
+++ b/Microbenchmarks/STREAM/dpu/copyw.c
@@ -18,12 +18,16 @@ __host dpu_results_t DPU_RESULTS[NR_TASKLETS];
 
 // Copy
 static void copyw_dpu(T *bufferB, T *bufferA) {
-
+#if UNROLL
     #pragma unroll
     for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
         bufferB[i] = bufferA[i];
     }
-
+#else
+    for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
+        bufferB[i] = bufferA[i];
+    }
+#endif
 }
 
 // Barrier
diff --git a/Microbenchmarks/STREAM/dpu/scale.c b/Microbenchmarks/STREAM/dpu/scale.c
index 4247aac..53a70a2 100644
--- a/Microbenchmarks/STREAM/dpu/scale.c
+++ b/Microbenchmarks/STREAM/dpu/scale.c
@@ -18,12 +18,16 @@ __host dpu_results_t DPU_RESULTS[NR_TASKLETS];
 
 // Scale
 static void scale_dpu(T *bufferB, T *bufferA, T scalar) {
-
+#if UNROLL
     #pragma unroll
     for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
         bufferB[i] = scalar * bufferA[i];
     }
-
+#else
+    for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
+        bufferB[i] = scalar * bufferA[i];
+    }
+#endif
 }
 
 // Barrier
diff --git a/Microbenchmarks/STREAM/dpu/triad.c b/Microbenchmarks/STREAM/dpu/triad.c
index e81cb59..9e641ea 100644
--- a/Microbenchmarks/STREAM/dpu/triad.c
+++ b/Microbenchmarks/STREAM/dpu/triad.c
@@ -18,12 +18,16 @@ __host dpu_results_t DPU_RESULTS[NR_TASKLETS];
 
 // Triad
 static void triad_dpu(T *bufferC, T *bufferA, T *bufferB, T scalar) {
-
+#if UNROLL
     #pragma unroll
     for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
         bufferC[i] = bufferA[i] + scalar * bufferB[i];
     }
-
+#else
+    for (unsigned int i = 0; i < BLOCK_SIZE / sizeof(T); i++){
+        bufferC[i] = bufferA[i] + scalar * bufferB[i];
+    }
+#endif
 }
 
 // Barrier