Update example Energibridge run_data population.

All the energy related columns of the energibridge output represent cumulative values that RAPL, for example, publishes.

Because of that, in order to save the total energy consumed by all parts of the system during the run we need to take the difference between the last and first values in each column.

Author: andrei-calin-dragomir

examples/energibridge-profiling/RunnerConfig.py

@@ -123,10 +123,10 @@ def populate_run_data(self, context: RunnerContext) -> Optional[Dict[str, Any]]:
         # energibridge.csv - Power consumption of the whole system
         df = pd.read_csv(context.run_dir / f"energibridge.csv")
         run_data = {
-                'dram_energy'   : round(df['DRAM_ENERGY (J)'].sum(), 3),
-                'package_energy': round(df['PACKAGE_ENERGY (J)'].sum(), 3),
-                'pp0_energy'    : round(df['PP0_ENERGY (J)'].sum(), 3),
-                'pp1_energy'    : round(df['PP1_ENERGY (J)'].sum(), 3),
+            'dram_energy': round(df['DRAM_ENERGY (J)'].iloc[-1] - df['DRAM_ENERGY (J)'].iloc[0], 3),
+            'package_energy': round(df['PACKAGE_ENERGY (J)'].iloc[-1] - df['PACKAGE_ENERGY (J)'].iloc[0], 3),
+            'pp0_energy': round(df['PP0_ENERGY (J)'].iloc[-1] - df['PP0_ENERGY (J)'].iloc[0], 3),
+            'pp1_energy': round(df['PP1_ENERGY (J)'].iloc[-1] - df['PP1_ENERGY (J)'].iloc[0], 3),
         }
         return run_data