Constantly getting Error "pynvml.nvml.NVMLError_NotSupported: Not Supported"

[KamalBoolchandani]

Hello, trying to use codecarbon on windows 10 with python version 3.10 with latest necessary pip packages but still seeing the error
""pynvml.nvml.NVMLError_NotSupported: Not Supported"
I have nvidia GeForce 930mx graphic card
here is the sample code I am trying to run
from codecarbon import EmissionsTracker

    n = 1000**7
    return n * (n + 1) // 2
def foo():
    # Start CodeCarbon tracker
    tracker = EmissionsTracker()

    try:
        # Start measuring emissions
        tracker.start()

        # Run power-intensive task
        result = power_intensive_task()
        print("Result:", result)

    finally:
        # Stop measuring emissions
        tracker.stop()

        # Print the emissions report
        print("Carbon Emissions:", tracker._emissions)

if __name__ == "__main__":
    foo()

and here is complete error

[codecarbon INFO @ 23:58:24] [setup] RAM Tracking...
[codecarbon INFO @ 23:58:24] [setup] GPU Tracking...
[codecarbon INFO @ 23:58:24] Tracking Nvidia GPU via pynvml
[codecarbon DEBUG @ 23:58:24] GPU available. Starting setup
Traceback (most recent call last):
  File "c:\Users\kamal\OneDrive\Desktop\Hackathon2023\testcode\code-carbon.py", line 26, in <module>
    food()
  File "c:\Users\kamal\OneDrive\Desktop\Hackathon2023\testcode\code-carbon.py", line 8, in food
    tracker = EmissionsTracker()
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\emissions_tracker.py", line 284, in __init__
    gpu_devices = GPU.from_utils(self._gpu_ids)
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\external\hardware.py", line 119, in from_utils
    return cls(gpu_ids=gpu_ids)
  File "<string>", line 4, in __init__
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\external\hardware.py", line 61, in __post_init__
    self.devices = AllGPUDevices()
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 186, in __init__
    gpu_device = GPUDevice(handle=handle, gpu_index=i)
  File "<string>", line 8, in __init__
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 24, in __post_init__
    self.last_energy = self._get_energy_kwh()
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 28, in _get_energy_kwh
    return Energy.from_millijoules(self._get_total_energy_consumption())
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 95, in _get_total_energy_consumption
    return pynvml.nvmlDeviceGetTotalEnergyConsumption(self.handle)
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 2411, in nvmlDeviceGetTotalEnergyConsumption
    _nvmlCheckReturn(ret)
  File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 833, in _nvmlCheckReturn
    raise NVMLError(ret)
pynvml.nvml.NVMLError_NotSupported: Not Supported

already tried to use decorator @trackemissions as well
TIY

[LuisBlanche]

Hi, thanks for using CodeCarbon ! Quick question : are you able to run the command nvidia-smi from the command line ?

[KamalBoolchandani]

sure , also i had updated to latest drivers avaialble
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 546.17 Driver Version: 546.17 CUDA Version: 12.3 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name TCC/WDDM | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce 930MX WDDM | 00000000:01:00.0 Off | N/A |
| N/A 0C P0 N/A / 200W | 0MiB / 2048MiB | 1% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+

+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| No running processes found |
+---------------------------------------------------------------------------------------+

[LuisBlanche]

Are you able to run the following code ?

import pynvml
handle = pynvml.nvmlDeviceGetHandleByIndex(0) 
pynvml.nvmlDeviceGetTotalEnergyConsumption(handle)

If not then the issue is probably with pynvml or something with your installation .
Unfortunately, it is difficult to reproduce the bug since we would need the same GPU available.

[KamalBoolchandani]

unfortunatly not i am still seeing more or less sam e error
handle = pynvml.nvmlDeviceGetHandleByIndex(0)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 1977, in nvmlDeviceGetHandleByIndex
fn = _nvmlGetFunctionPointer("nvmlDeviceGetHandleByIndex_v2")
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 848, in _nvmlGetFunctionPointer
raise NVMLError(NVML_ERROR_UNINITIALIZED)
pynvml.nvml.NVMLError_Uninitialized: Uninitialized

[LuisBlanche]

My mistake sorry pynvml was not initialized in my previous snippet
can you try:

import pynvml
pynvml.nvmlInit()
device_count = pynvml.nvmlDeviceGetCount()
handle = pynvml.nvmlDeviceGetHandleByIndex(0) 
pynvml.nvmlDeviceGetTotalEnergyConsumption(handle)

[KamalBoolchandani]

downgraded to codecarbon2.2.5 and got some more insights
[codecarbon INFO @ 01:36:40] [setup] RAM Tracking...
[codecarbon INFO @ 01:36:40] [setup] GPU Tracking...
[codecarbon INFO @ 01:36:40] Tracking Nvidia GPU via pynvml
[codecarbon DEBUG @ 01:36:40] Failed to retrieve gpu information
Traceback (most recent call last):
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 190, in get_gpu_details
"power_usage": get_power_usage(handle),
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 68, in get_power_usage
return pynvml.nvmlDeviceGetPowerUsage(handle)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 2404, in nvmlDeviceGetPowerUsage
_nvmlCheckReturn(ret)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 833, in _nvmlCheckReturn
raise NVMLError(ret)
pynvml.nvml.NVMLError_NotSupported: Not Supported
[codecarbon INFO @ 01:36:40] [setup] CPU Tracking...
[codecarbon DEBUG @ 01:36:40] Not using PowerGadget, an exception occurred while instantiating IntelPowerGadget : Intel Power Gadget executable not found on win32
[codecarbon DEBUG @ 01:36:40] Not using the RAPL interface, an exception occurred while instantiating IntelRAPL : Platform not supported by Intel RAPL Interface
[codecarbon WARNING @ 01:36:40] No CPU tracking mode found. Falling back on CPU constant mode.
[codecarbon DEBUG @ 01:36:42] CPU : We detect a Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz with a TDP of 15 W
[codecarbon INFO @ 01:36:42] CPU Model on constant consumption mode: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
[codecarbon DEBUG @ 01:36:42] Failed to retrieve gpu information
Traceback (most recent call last):
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 190, in get_gpu_details
"power_usage": get_power_usage(handle),
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 68, in get_power_usage
return pynvml.nvmlDeviceGetPowerUsage(handle)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 2404, in nvmlDeviceGetPowerUsage
_nvmlCheckReturn(ret)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 833, in _nvmlCheckReturn
raise NVMLError(ret)
pynvml.nvml.NVMLError_NotSupported: Not Supported
[codecarbon INFO @ 01:36:42] >>> Tracker's metadata:
[codecarbon INFO @ 01:36:42] Platform system: Windows-10-10.0.19045-SP0
[codecarbon INFO @ 01:36:42] Python version: 3.10.11
[codecarbon INFO @ 01:36:42] CodeCarbon version: 2.2.5
[codecarbon INFO @ 01:36:42] Available RAM : 11.889 GB
[codecarbon INFO @ 01:36:42] CPU count: 8
[codecarbon INFO @ 01:36:42] CPU model: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
[codecarbon INFO @ 01:36:42] GPU count: 1
[codecarbon INFO @ 01:36:42] GPU model: 1 x NVIDIA GeForce 930MX
[codecarbon DEBUG @ 01:36:43] Not running on AWS
[codecarbon DEBUG @ 01:36:44] Not running on Azure
[codecarbon DEBUG @ 01:36:45] Not running on GCP
[codecarbon INFO @ 01:36:46] ApiClient Successfully registered your run on the API.

Run ID: 1053ff0c-e0d3-4f4a-81d0-175423f74908
Experiment ID: 5b45b0fe-075a-4543-8abb-48e906b854ff

Result: 500000000000000000000500000000000000000000
[codecarbon INFO @ 01:36:46] Energy consumed for RAM : 0.000000 kWh. RAM Power : 4.458486557006836 W
[codecarbon DEBUG @ 01:36:46] RAM : 4.46 W during 0.00 s [measurement time: 0.0050]
[codecarbon DEBUG @ 01:36:46] Failed to retrieve gpu information
Traceback (most recent call last):
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 190, in get_gpu_details
"power_usage": get_power_usage(handle),
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 68, in get_power_usage
return pynvml.nvmlDeviceGetPowerUsage(handle)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 2404, in nvmlDeviceGetPowerUsage
_nvmlCheckReturn(ret)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 833, in _nvmlCheckReturn
raise NVMLError(ret)
pynvml.nvml.NVMLError_NotSupported: Not Supported
[codecarbon INFO @ 01:36:46] Energy consumed for all GPUs : 0.000000 kWh. Total GPU Power : 0.0 W
[codecarbon DEBUG @ 01:36:46] Failed to retrieve gpu information
Traceback (most recent call last):
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 190, in get_gpu_details
"power_usage": get_power_usage(handle),
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\codecarbon\core\gpu.py", line 68, in get_power_usage
return pynvml.nvmlDeviceGetPowerUsage(handle)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 2404, in nvmlDeviceGetPowerUsage
_nvmlCheckReturn(ret)
File "C:\Users\kamal\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pynvml\nvml.py", line 833, in _nvmlCheckReturn
raise NVMLError(ret)
pynvml.nvml.NVMLError_NotSupported: Not Supported
[codecarbon DEBUG @ 01:36:46] GPU : 0.00 W during 0.01 s [measurement time: 0.0050]
[codecarbon INFO @ 01:36:46] Energy consumed for all CPUs : 0.000000 kWh. Total CPU Power : 7.5 W
[codecarbon DEBUG @ 01:36:46] CPU : 7.50 W during 0.02 s [measurement time: 0.0000]
[codecarbon INFO @ 01:36:46] 0.000000 kWh of electricity used since the beginning.
[codecarbon DEBUG @ 01:36:46] last_duration=0.020548582077026367

[codecarbon DEBUG @ 01:36:46] We apply an energy mix of 637 g.CO2eq/kWh for India
[codecarbon DEBUG @ 01:36:46] EmissionsData(timestamp='2023-11-30T01:36:46', project_name='codecarbon', run_id='1053ff0c-e0d3-4f4a-81d0-175423f74908', duration=0.02254509925842285, emissions=2.8852795252481015e-08, emissions_rate=1.2797812474345885e-06, cpu_power=7.5, gpu_power=0.0, ram_power=4.458486557006836, cpu_energy=4.280954599380493e-08, gpu_energy=0.0, ram_energy=2.4723283411276498e-09, energy_consumed=4.5281874334932584e-08, country_name='India', country_iso_code='IND', region='maharashtra', cloud_provider='', cloud_region='', os='Windows-10-10.0.19045-SP0', python_version='3.10.11', codecarbon_version='2.2.5', cpu_count=8, cpu_model='Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz', gpu_count=1, gpu_model='1 x NVIDIA GeForce 930MX', longitude=73.7286, latitude=18.6161, ram_total_size=11.889297485351562, tracking_mode='machine', on_cloud='N', pue=1.0)
[codecarbon DEBUG @ 01:36:46] We apply an energy mix of 637 g.CO2eq/kWh for India
[codecarbon DEBUG @ 01:36:46] EmissionsData(timestamp='2023-11-30T01:36:46', project_name='codecarbon', run_id='1053ff0c-e0d3-4f4a-81d0-175423f74908', duration=0.037804603576660156, emissions=2.8852795252481015e-08, emissions_rate=7.632085122642096e-07, cpu_power=7.5, gpu_power=0.0, ram_power=4.458486557006836, cpu_energy=4.280954599380493e-08, gpu_energy=0.0, ram_energy=2.4723283411276498e-09, energy_consumed=4.5281874334932584e-08, country_name='India', country_iso_code='IND', region='maharashtra', cloud_provider='', cloud_region='', os='Windows-10-10.0.19045-SP0', python_version='3.10.11', codecarbon_version='2.2.5', cpu_count=8, cpu_model='Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz', gpu_count=1, gpu_model='1 x NVIDIA GeForce 930MX', longitude=73.7286, latitude=18.6161, ram_total_size=11.889297485351562, tracking_mode='machine', on_cloud='N', pue=1.0)
[codecarbon WARNING @ 01:36:46] ApiClient : emissions not sent because of a duration smaller than 1.
Carbon Emissions: <codecarbon.core.emissions.Emissions object at 0x0000021C9BF45720>

for the sample code

def power_intensive_task():
    n = 1000**7
    return n * (n + 1) // 2


def track():
    # Start CodeCarbon tracker
    tracker = EmissionsTracker()

    try:
        # Start measuring emissions
        tracker.start()

        # Run power-intensive task
        result = power_intensive_task()
        print("Result:", result)

    finally:
        # Stop measuring emissions
        tracker.stop()

        # Print the emissions report
        print("Carbon Emissions:", tracker._emissions)

if __name__ == "__main__":
    track()```