diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 65e62e326f05460762d095585d9f663c032ba7a2..7a159bdd865e88d1e89193743c964ebbda6ad626 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -26,6 +26,7 @@ test:
   image: i10git.cs.fau.de:5005/pycodegen/pystencils-benchmark/pystencils-benchmark
   tags:
     - docker
+    - cuda
   script:
     - pip install tox
     - echo $TOX_ENV
diff --git a/Dockerfile b/Dockerfile
index 2b8f8a89b16b3fe04173260d3b026dc2b5787324..dcb5493f7ed1d3f895340ad38b2182d1a0cb52c7 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,7 +1,7 @@
-FROM ubuntu:latest
+FROM nvidia/cuda:12.1.1-devel-ubuntu22.04
 
 LABEL maintainer="jan.hoenig@fau.de"
-LABEL version="0.1"
+LABEL version="0.2"
 LABEL description="Custom docker image for pystencils-benchmark"
 
 ARG DEBIAN_FRONTEND=noninteractive
diff --git a/tests/test_benchmark.py b/tests/test_benchmark.py
index 929d4f8a59d2fec219fb948da9c87e8ff9c31c2f..62881eccc8f9a7f7f6232bce9a2f17b939313c86 100755
--- a/tests/test_benchmark.py
+++ b/tests/test_benchmark.py
@@ -57,5 +57,5 @@ def test_generate_gpu(kwargs):
     with tempfile.TemporaryDirectory(dir=Path.cwd()) as temp_dir:
         temp_dir = Path(temp_dir)
         pb.gpu.generate_benchmark(kernel_vadd, temp_dir, compiler=compiler, **kwargs)
-        # subprocess.run(['make', '-C', f'{temp_dir}'], check=True)
-        # subprocess.run([f'{temp_dir}/benchmark-{compiler.name}', '10'], check=True)
+        subprocess.run(['make', '-C', f'{temp_dir}'], check=True)
+        subprocess.run([f'{temp_dir}/benchmark-{compiler.name}', '10'], check=True)