Cleaned up code

vecadd_cpu.cpp

@@ -9,7 +9,7 @@ void add(int n, float *x, float *y) {
 }
 
 int main(void) {
-  int N = 1 << 29; // Setting problem size to 1<<29 (536,870,912 elements)
+  int N = 1 << 29;
 
   float *x = new float[N];
   float *y = new float[N];
@@ -20,10 +20,10 @@ int main(void) {
     y[i] = 2.0f;
   }
 
-  // Timer starts before the add function call
+  // Start chrono timer
   auto start_time = std::chrono::high_resolution_clock::now();
 
-  // Run kernel on the elements on the CPU
+  // Run kernel on N << 29 elements on the CPU
   add(N, x, y);
 
   // Timer ends after the add function call

vecadd_gpu_1t.cu

@@ -1,6 +1,6 @@
-#include <iostream>
 #include <cmath>
 #include <cuda_runtime.h>
+#include <iostream>
 
 // function to add the elements of two arrays
 __global__ void add(int n, float *x, float *y) {
@@ -9,7 +9,7 @@ __global__ void add(int n, float *x, float *y) {
 }
 
 int main(void) {
-  int N = 1 << 29; // Setting problem size to 1<<29 (536,870,912 elements)
+  int N = 1 << 29;
 
   float *x, *y;
 
@@ -40,4 +40,4 @@ int main(void) {
   cudaFree(y);
 
   return 0;
-}
+}

vecadd_gpu_256t.cu

@@ -1,6 +1,6 @@
-#include <iostream>
 #include <cmath>
 #include <cuda_runtime.h>
+#include <iostream>
 
 // function to add the elements of two arrays
 __global__ void add(int n, float *x, float *y) {
@@ -11,7 +11,7 @@ __global__ void add(int n, float *x, float *y) {
 }
 
 int main(void) {
-  int N = 1 << 29; // Setting problem size to 1<<29 (536,870,912 elements)
+  int N = 1 << 29;
 
   float *x, *y;
 
@@ -42,4 +42,4 @@ int main(void) {
   cudaFree(y);
 
   return 0;
-}
+}

vecadd_gpu_256t_mb.cu

@@ -1,6 +1,6 @@
-#include <iostream>
 #include <cmath>
 #include <cuda_runtime.h>
+#include <iostream>
 
 // function to add the elements of two arrays
 __global__ void add(int n, float *x, float *y) {
@@ -11,7 +11,7 @@ __global__ void add(int n, float *x, float *y) {
 }
 
 int main(void) {
-  int N = 1 << 29; // Setting problem size to 1<<29 (536,870,912 elements)
+  int N = 1 << 29;
 
   float *x, *y;
 
@@ -30,7 +30,6 @@ int main(void) {
   // Number of blocks in the grid
   int numberOfBlocks = (N + threadsPerBlock - 1) / threadsPerBlock;
 
-  // Print out the number of thread blocks
   std::cout << "Number of thread blocks: " << numberOfBlocks << std::endl;
 
   // Run kernel on the elements on the GPU with multiple blocks and threads
@@ -50,4 +49,4 @@ int main(void) {
   cudaFree(y);
 
   return 0;
-}
+}

vecadd_gpu_256t_mb_prefetch.cu

@@ -1,6 +1,6 @@
-#include <iostream>
 #include <cmath>
 #include <cuda_runtime.h>
+#include <iostream>
 
 // function to add the elements of two arrays
 __global__ void add(int n, float *x, float *y) {
@@ -11,7 +11,7 @@ __global__ void add(int n, float *x, float *y) {
 }
 
 int main(void) {
-  int N = 1 << 29; // Setting problem size to 1<<29 (536,870,912 elements)
+  int N = 1 << 29;
 
   float *x, *y;
 
@@ -35,7 +35,6 @@ int main(void) {
   // Number of blocks in the grid
   int numberOfBlocks = (N + threadsPerBlock - 1) / threadsPerBlock;
 
-  // Print out the number of thread blocks
   std::cout << "Number of thread blocks: " << numberOfBlocks << std::endl;
 
   // Run kernel on the elements on the GPU with multiple blocks and threads
@@ -55,4 +54,4 @@ int main(void) {
   cudaFree(y);
 
   return 0;
-}
+}