feat: Initial Commit

.clangd

@@ -0,0 +1,6 @@
+CompileFlags:
+  Add:
+    - --cuda-path=/opt/cuda
+    - --cuda-gpu-arch=sm_89   # Replace XX with your actual GPU architecture, e.g., 86
+    - -I/opt/cuda/include
+    - -L/opt/cuda/lib64

main.cu

@@ -0,0 +1,41 @@
+#include <stdio.h>
+
+__global__ void cuda_hello(int* number, int* out) {
+    *out = *number;
+    for(int i = 0; i < 255; i++) {
+        *out += 1;
+    }
+}
+
+int main() {
+    int *number, *out;
+    int *d_number, *d_out;
+    number = (int*)malloc(sizeof(int));
+    out = (int*)malloc(sizeof(int));
+
+    *number = 22;
+    *out = 0;
+
+    cudaMalloc((void**)&d_number, sizeof(int));
+    cudaMemcpy(d_number, number, sizeof(int), cudaMemcpyHostToDevice);
+    cudaMalloc((void**)&d_out, sizeof(int));
+    cudaMemcpy(d_out, out, sizeof(int), cudaMemcpyHostToDevice);
+
+
+    cuda_hello<<<1,1>>>(d_number, d_out);
+
+    cudaDeviceSynchronize();
+
+    cudaMemcpy(out, d_out, sizeof(int), cudaMemcpyDeviceToHost);
+
+    cudaFree(d_number);
+    cudaFree(d_out);
+
+
+    printf("Number: %i\n", *out);
+
+    free(number);
+    free(out);
+
+    return 0;
+}