diff --git a/README.md b/README.md
index 2f04d66a9e1e06bb0c50cad095897056214e32c9..515ff8db7fb1da6e479c1f99ae7a0501170f96ef 100644
--- a/README.md
+++ b/README.md
@@ -8,7 +8,7 @@ The request to the database also allows a statistical analysis of the requested
 The mean and standard deviation of all stations within a cell are computed. 
 
 The tool handles the request to the database over the REST API and the subsequent processing.
-The results are provided as xarray objects for subsequent processing by the user.
+The results of the gridding are provided as xarray objects for subsequent processing by the user.
 
 This project is in beta with the intended basic functionalities. 
 The documentation is work in progress.
@@ -70,6 +70,7 @@ quality_controll.ipynb
 ```
 Notebook for downloading and visualization of data. 
 The data are downloaded and reused for subsequent executions of this notebook.
+The gridding is done on the downloaded data. Gridded data are not saved to disc. 
 
 # Supported Grids
 
diff --git a/tests/benchmark.py b/tests/benchmark.py
index 7390b1751e9dbddc39eaf658d303798cef4aa598..71fb5b4e2ea2f4629c91adfe1db621952483a6ad 100644
--- a/tests/benchmark.py
+++ b/tests/benchmark.py
@@ -16,6 +16,8 @@ TEST_ROOT = Path(__file__).parent
 
 
 def get_toar_response(start, end):
+    """manual request to the TOAR database
+    """
     end_with_padding = end + timedelta(1)
 
     response = requests.get(
@@ -35,6 +37,8 @@ def get_toar_response(start, end):
 
 
 def wait_for_data(response):
+    """waiting for data of a manual request
+    """
     tries = 0
     while True:
         print(f"n tries: {tries}")
@@ -53,6 +57,8 @@ def wait_for_data(response):
 
 
 def time_toar_response(start, end):
+    """benchmark with manual creation of request to TOAR database
+    """
     print(f"starting request for {start}-{end}")
 
     timer_start = datetime.now()
@@ -64,6 +70,8 @@ def time_toar_response(start, end):
 
 
 def wait_for_client_response(client, sample):
+    """waiting for response. get_data throws exception after 30min, if results are not available.
+    """
     for half_hours in range(20):
         try:
             client.get_data(sample)
@@ -74,10 +82,15 @@ def wait_for_client_response(client, sample):
 def time_rest_client_response(start, end):
     print(f"starting request for {start}-{end}")
 
+    path_cache = TEST_ROOT / "temp_data_cache"
+    path_data  = TEST_ROOT / "temp_data"
+    path_cache.mkdir(parents=True, exist_ok=True)
+    path_data.mkdir(parents=True, exist_ok=True)
+
     rest_client = AnalysisServiceDownload(
         STATISTICS_ENDPOINT,
-        TEST_ROOT / "temp_data_cache",
-        TEST_ROOT / "temp_data",
+        path_cache,
+        path_data,
         use_downloaded=False,
     )
     time_window = TimeSample(start, end, SAMPLING)