Cleanup: Refactor benchmark table.

benchmarks/README.md

@@ -2,20 +2,20 @@
 
 ## Run
 
-Run the `run_all.py` script to run all benchmarks.
+If you want to evaluate a single benchmark, adopt the `run_single.py` script
+and (within the `benchmarks` directory) run
 
 ```bash
-python run_all.py
+python run_single.py
 ```
 
-If you only want to evaluate a single benchmark, adopt the `run_single.py` script
-and run
+To run a specific benchmark, e.g., the sine benchmarks, run
 
 ```bash
-python run_single.py
+python sine/run_all.py
 ```
 
-## Visualize
+## MLFlow
 
 We use MLFlow to log the benchmark runs and you can run
 
@@ -25,8 +25,11 @@ mlflow ui
 
 to visualize the logged experiments in your browser.
 
-In order to visualize the best runs for every benchmark in an HTML table, run
+## Documentation
+
+In order to process the data of the best runs for the benchmark page in
+the documentation, run
 
 ```bash
-python process.py
+python results/process.py
 ```