diff --git a/README.md b/README.md
new file mode 100644
index 0000000..0000fb0
--- /dev/null
+++ b/README.md
@@ -0,0 +1,8 @@
+
+Example cuda code compilation in nix.
+
+```bash
+$ nix run github:osbm/cuda-kernel-nix
+Free mem: 22112 MB Total mem: 24078 MB
+Free mem: 21 GB Total mem: 23 GB
+```
diff --git a/flake.lock b/flake.lock
index 157731b..e9703d4 100644
--- a/flake.lock
+++ b/flake.lock
@@ -2,11 +2,11 @@
   "nodes": {
     "nixpkgs": {
       "locked": {
-        "lastModified": 1734988233,
-        "narHash": "sha256-Ucfnxq1rF/GjNP3kTL+uTfgdoE9a3fxDftSfeLIS8mA=",
+        "lastModified": 1756159630,
+        "narHash": "sha256-ohMvsjtSVdT/bruXf5ClBh8ZYXRmD4krmjKrXhEvwMg=",
         "owner": "nixos",
         "repo": "nixpkgs",
-        "rev": "de1864217bfa9b5845f465e771e0ecb48b30e02d",
+        "rev": "84c256e42600cb0fdf25763b48d28df2f25a0c8b",
         "type": "github"
       },
       "original": {
diff --git a/flake.nix b/flake.nix
index fabb8b0..230130d 100644
--- a/flake.nix
+++ b/flake.nix
@@ -29,21 +29,40 @@
     packages."${system}" = {
       # just one main.cu file for now
       # compile it and place it to $out/bin/cuda-kernel
-      cuda-kernel = pkgs.stdenv.mkDerivation {
-        name = "cuda-kernel";
-        src = ./.;
-        buildInputs = with pkgs; [
-          cudatoolkit
-          cudaPackages.cuda_cudart
-          linuxPackages.nvidia_x11
-        ];
-        LD_LIBRARY_PATH = "${pkgs.linuxPackages.nvidia_x11}/lib";
-        buildPhase = ''
-          nvcc -o main main.cu
-          mkdir -p $out/bin
-          mv main $out/bin/cuda-kernel
-        '';
-      };
+      cuda-kernel = let
+        # Use CUDA 12.0 which should be compatible with driver 580.76.05
+        cudaPackages = pkgs.cudaPackages;
+        cudatoolkit = cudaPackages.cudatoolkit;
+      in
+        cudaPackages.backendStdenv.mkDerivation {
+          name = "cuda";
+          src = ./main.cu;
+          unpackPhase = "true";
+          buildInputs = [
+            cudatoolkit
+            cudaPackages.cuda_cudart
+          ];
+          nativeBuildInputs = [
+            cudatoolkit
+            pkgs.autoAddDriverRunpath
+          ];
+          LD_LIBRARY_PATH = "${cudatoolkit}/lib:${cudaPackages.cuda_cudart}/lib";
+          buildPhase = ''
+            echo "GCC Version:"
+            gcc --version
+            echo "NVCC Version:"
+            nvcc --version
+
+            echo "Building CUDA program..."
+            nvcc -arch=sm_89 -o main $src
+            mkdir -p $out/bin
+            cp main $out/bin/
+          '';
+          meta = {
+            mainProgram = "main";
+            platforms = ["x86_64-linux"];
+          };
+        };
       default = self.packages."${system}".cuda-kernel;
     };
   };
diff --git a/main.cu b/main.cu
index d99fd87..6ce6ea2 100644
--- a/main.cu
+++ b/main.cu
@@ -1,13 +1,13 @@
-#include <unistd.h>
 #include <iostream>
+#include "cuda.h"
 
-
-int main(){
-
-    size_t mf, ma;
-    cudaMemGetInfo(&mf, &ma);
-    std::cout << "Free memory (mb): " << mf/1024/1024 << std::endl;
-    std::cout << "Total memory (mb): " << ma/1024/1024 << std::endl;
-
-    return 0;
+int main() {
+  size_t free_t, total_t;
+  cudaError_t err;
+  err = cudaMemGetInfo(&free_t, &total_t);
+  if (err != cudaSuccess) { printf("cudaMalloc d_steps_arr error: %s\n", cudaGetErrorString(err)); return 1; }
+  std::cout << "Free mem: " << free_t / (1024 * 1024) << " MB Total mem: " << total_t / (1024 * 1024) << " MB" << std::endl;
+  std::cout << "Free mem: " << free_t / (1024 * 1024 * 1024) << " GB Total mem: " << total_t / (1024 * 1024 * 1024) << " GB" << std::endl;
+  return 0;
 }
+