Run coverage combine, even if there is no coverage

As it does not raise an error.
Therefore, the benchmark and cupy-tests can be combined in a single job.

.gitlab-ci.yml

@@ -40,7 +40,7 @@ test_JUWELS:
     - shell
   parallel:
     matrix:
-      - SHELL_SCRIPT: [ 'cupy' ]
+      - SHELL_SCRIPT: ['benchmark', 'cupy' ]
   artifacts:
     when: always
     paths:
@@ -76,51 +76,6 @@ test_JUWELS:
     - ls -lh .
 
 
-benchmark_JUWELS:
-  stage: benchmark
-  rules:
-    - if: $CI_COMMIT_MESSAGE !~ /.*\[CI-no-benchmarks\]/
-  tags:
-    - jacamar
-    - juwels
-    - login
-    - shell
-  parallel:
-    matrix:
-      - SHELL_SCRIPT: [ 'benchmark']
-  artifacts:
-    when: always
-    paths:
-      - coverage_*.dat
-      - sbatch.err
-      - sbatch.out
-  before_script:
-    - mkdir -p benchmarks
-    # load the latest Python module (currently 3.11)
-    - module --force purge
-    - module load Stages/2024
-    - module load GCC
-    - module load OpenMPI
-    - module load FFTW
-    - module load mpi4py
-    - module load SciPy-Stack
-    - module load CuPy
-    - pwd
-    - ls -lah
-    - pip install -e .
-    - pip install pytest-benchmark
-  script:
-    # - touch benchmarks/output.json
-    - echo $SYSTEMNAME
-    - sbatch --wait etc/juwels_${SHELL_SCRIPT}.sh
-  after_script:
-    - echo "Following Errors occured:"
-    - cat sbatch.err
-    - echo "Following was written to stdout:"
-    - cat sbatch.out
-    - ls -lh .
-
-
 #test_kit:
 #  image: rcaspart/micromamba-cuda
 #  stage: benchmark