ProcessC is a comprehensive program designed to monitor energy usage and carbon emissions for running process-based modeling simulations. It provides detailed tracking and analysis of CPU, GPU, and RAM power consumption while calculating total carbon emissions based on country or regional grid carbon intensity.
ProcessC now offers smart computation grid algorithms based on the computation power of designated CPU/GPU requirements, optimizing energy efficiency for your simulations.
- Multi-Component Monitoring: Tracks CPU, GPU, and RAM energy usage in real-time
- Carbon Footprint Calculation: Computes total carbon emissions for running simulations
- Cross-Platform CPU Support: Compatible with both Intel and AMD processors
- Flexible Project Management: Support for existing projects or new project creation
- Multiple Data Sources: Fetches carbon intensity data from various reliable sources
- Smart Grid Algorithms: Optimizes computation based on hardware requirements
- Comprehensive Logging: Detailed CSV output for analysis and reporting
- Python: 3.x or higher
- Required Libraries:
os, json, datetime, pprint, tabulate, requests, pandas, psutil, cpuinfo, GPUtil, logging, csv, art
-
Clone the repository:
git clone https://github.com/lzwei196/ProcessC.git cd ProcessC -
Install dependencies:
pip install -r requirements.txt
-
Ensure configuration: Make sure
conf.jsonis available in the working directory.
Before running ProcessC, ensure you have:
- CLI Command: Your process-based model must be executable from command line (required for bash mode)
- Process Name: The exact name of your program as it appears in system processes
- Model Directory: The working directory path of your model
Launch ProcessC with:
python main.pyProcessC automatically checks for existing projects in conf.json. If none are found, you'll be guided through creating a new project:
-
Project Setup:
- Enter project name
- Choose monitoring mode (Bash or Direct)
- Configure mode-specific parameters
-
System Configuration:
- Auto-detection of CPU and GPU specifications
- Manual input option for custom configurations
- Grid carbon intensity data source selection
-
Save Configuration:
- Project settings saved for future use
- Reusable configurations for similar simulations
Once configured, ProcessC will:
- Real-time Monitoring: Continuously track energy consumption
- Data Logging: Record CPU, GPU, and RAM power usage
- Emission Calculation: Compute total energy usage and carbon footprint
- Results Output: Generate tabular results and CSV files
| Metric | Value |
|---|---|
| Project Name | MyProject |
| Elapsed Time (seconds) | 3,600 |
| CPU Energy (kWh) | 0.05 |
| GPU Energy (kWh) | 0.03 |
| RAM Power Usage (kWh) | 0.02 |
| Total Energy Usage (kWh) | 0.10 |
| Grid Carbon Intensity (gCO₂/kWh) | 500.0 |
| Total Carbon Emission (gCO₂) | 50.0 |
ProcessC can serve as a wrapper for auto-calibration processes:
- Currently supports RS-DPCF auto-calibration
- Extensible architecture for other auto-calibration software
- Contact us for additional software support
- Automatic System Detection: CPU, GPU, and RAM specifications
- Connectivity Verification: Internet connection status checking
- Location Services: Automatic region and country detection
- Comprehensive Logging: All monitoring data saved to CSV files in
/outputfolder
- AMD CPUs on Windows: Cannot use Intel Power Gadget; uses AMD's default TDP values instead
- Direct Mode: ProcessC verifies if the target program is running
- CMD-Based Executables: May appear as "OpenConsole.log" - ensure no other programs with this name are running
- OurWorldInData: Local database (no internet required)
- ElectricityMap: Requires paid subscription or manual carbon intensity input for specific locations/years
If you use ProcessC in your research, please cite our paper:
@article{processc2024,
title={ProcessC: Energy and Carbon Emission Monitoring for Process-Based Modeling Simulations},
journal={Resources, Conservation and Recycling},
year={2024},
doi={10.1016/j.resconrec.2024.108101},
url={https://doi.org/10.1016/j.resconrec.2024.108101}
}Developers: Ziwei Li, Zhiming Qi, Birk Li, Junzeng Xu, Ruiqi Wu, Yuchen Liu
Affiliations:
- Qi Lab, McGill University, Bioresource Engineering
- Hohai University
For questions, support requests, or additional process-based model integration:
📧 Email: leo.li@mail.mcgill.ca
We welcome feedback and are happy to provide support for integrating additional process-based models.
This project is licensed under the MIT License - see the LICENSE file for details.
ProcessC - Making environmental impact monitoring accessible for computational research