Feat: Raise RAPL permission error with informative message (#111)

Co-authored-by: Jae-Won Chung <[email protected]>

capriccio/generate.py

@@ -9,7 +9,6 @@
     python capriccio/generate.py /path/to/sentiment140.json
 """
 
-
 import argparse
 import os
 import warnings

tests/test_monitor.py

@@ -43,9 +43,9 @@ def getTotalEnergyConsumption(self):
         """Returns the total energy consumption of the specified powerzone. Units: mJ."""
         return CpuDramMeasurement(
             cpu_mj=float(next(self.cpu_energy)),
-            dram_mj=float(next(self.dram_energy))
-            if self.dram_energy is not None
-            else None,
+            dram_mj=(
+                float(next(self.dram_energy)) if self.dram_energy is not None else None
+            ),
         )
 
     def supportsGetDramEnergyConsumption(self):
@@ -281,9 +281,11 @@ def assert_window_begin(name: str, begin_time: int):
             for i in range(len(monitor.cpu_indices))
         }
         assert monitor.measurement_states[name].dram_energy == {
-            i: pytest.approx((200 + 5 * (begin_time - 4)) / 1000.0)
-            if i % 2 == 0
-            else None
+            i: (
+                pytest.approx((200 + 5 * (begin_time - 4)) / 1000.0)
+                if i % 2 == 0
+                else None
+            )
             for i in range(0, len(monitor.cpu_indices), 2)
         }
         pynvml_mock.nvmlDeviceGetTotalEnergyConsumption.assert_has_calls(

zeus/device/cpu/rapl.py

@@ -148,6 +148,14 @@ def __init__(self, message: str) -> None:
         super().__init__(message)
 
 
+class ZeusRAPLPermissionError(ZeusBaseCPUError):
+    """Zeus GPU exception that wraps No Permission to perform GPU operation."""
+
+    def __init__(self, message: str) -> None:
+        """Intialize the exception object."""
+        super().__init__(message)
+
+
 class RAPLFile:
     """RAPL File class for each RAPL file.
 
@@ -169,6 +177,10 @@ def __init__(self, path: str) -> None:
                 self.last_energy = float(energy_file.read().strip())
         except FileNotFoundError as err:
             raise ZeusRAPLFileInitError("Error reading package energy") from err
+        except PermissionError as err:
+            raise cpu_common.ZeusCPUNoPermissionError(
+                "Can't read file due to permission error"
+            ) from err
         try:
             with open(
                 os.path.join(path, "max_energy_range_uj"), "r"

zeus/monitor/energy.py

@@ -15,7 +15,7 @@
 from zeus.utils.framework import sync_execution as sync_execution_fn
 from zeus.device import get_gpus, get_cpus
 from zeus.device.gpu.common import ZeusGPUInitError, EmptyGPUs
-from zeus.device.cpu.common import ZeusCPUInitError, EmptyCPUs
+from zeus.device.cpu.common import ZeusCPUInitError, ZeusCPUNoPermissionError, EmptyCPUs
 
 logger = get_logger(__name__)
 
@@ -186,6 +186,14 @@ def __init__(
             self.cpus = get_cpus()
         except ZeusCPUInitError:
             self.cpus = EmptyCPUs()
+        except ZeusCPUNoPermissionError as err:
+            if cpu_indices:
+                raise RuntimeError(
+                    "Root privilege is required to read RAPL metrics. See "
+                    "https://ml.energy/zeus/getting_started/#system-privileges "
+                    "for more information or disable CPU measurement by passing cpu_indices=[] to "
+                    "ZeusMonitor"
+                ) from err
 
         # Resolve GPU indices. If the user did not specify `gpu_indices`, use all available GPUs.
         self.gpu_indices = (