A comprehensive Python-based tool for analyzing CPU and GPU performance data from MSI Afterburner logs. This tool provides detailed visualizations and statistics for system monitoring data, perfect for gamers, overclockers, and system performance enthusiasts.
Make sure to enable logging parameters under monitoring section in the MSI Afterburner app and double check whether log history to file option is marked during recording process and unchecked to stop recording.
- Temperature Monitoring: Individual core temperatures and overall CPU temperature
- Usage Analysis: Per-core and overall CPU utilization with visual load indicators
- Frequency Tracking: CPU clock speeds for all cores with frequency analysis
- Power Consumption: Both percentage-based and watts-based power monitoring with efficiency zones
- Performance Metrics: Temperature, usage, power consumption, and fan speeds
- Memory Monitoring: VRAM usage analysis with statistical summaries
- Clock Speeds: Core and memory clock frequency tracking
- Gaming Performance: Framerate analysis with average FPS indicators
- Dual Memory Tracking: Separate analysis for GPU VRAM and System RAM
- Usage Patterns: Memory consumption over time with average usage lines
- Smart Scaling: Automatic y-axis scaling based on memory capacity
- Energy Consumption: Calculate power usage and estimated electricity costs
- Statistical Analysis: Mean, min, max, median, and standard deviation for all metrics
- Performance Zones: Visual indicators for idle, normal, and high-load states
- Data Export: Save processed data in CSV and Excel formats
pip install pandas numpy matplotlib scikit-learn seaborn pathlibInstall additional libraries below if you want to use PostgreSQL.
pip install SQLAlchemy psycopg2- Python 3.7+
- Compatible with Windows, macOS, and Linux
-
Generate MSI Afterburner logs:
- Open MSI Afterburner
- Go to Settings → Monitoring
- Enable desired metrics (CPU temp, GPU temp, usage, etc.)
- MSI Afterburner takes logs in .hml format but users simply need to convert it to
.txt.
-
Run the analysis tool:
python performance_analyzer.py
-
Enter your log file path when prompted or use the default path
-
Choose analysis options from the interactive menu
-
Export your txt file into PostgreSQL database using
aft_postgre.py.
1 - CPU Temperature Analysis Individual core and overall CPU temperatures
2 - CPU Usage Analysis Per-core and overall CPU utilization
3 - CPU Frequency Analysis CPU clock speeds and frequency patterns
4 - CPU Power Analysis Power consumption (% and Watts) with energy costs
5 - GPU Analysis (Complete) Comprehensive GPU performance metrics
6 - Framerate Analysis Gaming performance and FPS analysis
7 - Memory Analysis GPU VRAM and System RAM usage
8 - Show Statistics Detailed statistical summaries
9 - Save Data Export processed data
q - Quit Exit the application
from performance_analyzer import PerformanceAnalyzer
# Initialize analyzer
analyzer = PerformanceAnalyzer("path/to/your/afterburner_log.txt")
# Load and process data
if analyzer.load_data():
# Generate CPU temperature analysis
analyzer.plot_cpu_temperatures()
# Analyze GPU performance
analyzer.plot_gpu_analysis()
# Get statistical summary
stats = analyzer.get_statistics()
print(stats)The tool automatically processes MSI Afterburner .txt log files with the following supported metrics:
CPU Metrics:
- CPU temperature, CPU1-CPUn temperature
- CPU usage, CPU1-CPUn usage
- CPU clock, CPU1-CPUn clock
- CPU power, Power consumption
- For devices that have multiple GPUs; GPU1 usage, GPU1 temperature and GPU1 power options have been added, If they exist they will be drawn on graphs.
GPU Metrics:
- GPU temperature, GPU usage
- Memory usage (VRAM), Core clock, Memory clock
- Power, Power percent, Fan speed
- Framerate (FPS)
System Metrics:
- RAM usage, Memory usage
- Timestamps with automatic date/time parsing
# Default Afterburner log path (Windows)
DEFAULT_PATH = r"C:\Python_Works\py\afterburner-data-analyzer\examples\texts\HMW_MHW.txt"
# Output directory for saved data
OUTPUT_DIR = "output/"- Plot Styles: Automatic fallback between seaborn and matplotlib styles
- Color Schemes: Professional color palettes with performance zone indicators
- Export Formats: CSV and Excel export with UTF-8 encoding
- Statistical Metrics: Configurable statistical analysis parameters
- Multi-panel Dashboards: Comprehensive performance overviews
- Time-series Plots: Performance metrics over time with trend analysis
- Performance Zones: Color-coded efficiency and load indicators
- Statistical Overlays: Average lines and performance thresholds
- Apache Superset (optional): Possible for users who know how to implement their database into Apache Superset for further analysis, example is given below.
performance_data.csv Comma-separated values for data analysis
performance_data.xlsx Excel format with formatted columns
GPU Performance Statistics:
========================================
GPU usage GPU temperature Power Framerate
mean 65.34 72.45 145.67 89.23
min 12.45 45.67 78.90 45.67
max 98.76 84.32 198.45 144.50
median 67.89 71.23 142.34 92.15
std 18.45 8.92 25.67 15.78
CPU Power Statistics:
=========================
CPU Power Percentage (CPU power):
Mean: 106.79%
Min: 0.0%
Max: 174.18%
Std: 19.33%
CPU Power Consumption (Power):
Mean: 158.92W
Min: 52.78W
Max: 179.63W
Std: 10.85W
Session Duration: 0.32 hours
Estimated Energy Consumption: 0.0501 kWh
Estimated Cost (at TL 2.59/kWh): TL 0.1298
- NVIDIA: RTX 40/30/20/10 series, GTX 16/10 series
- AMD: RX 7000/6000/5000 series, Radeon VII, Vega series
- Focus: Discrete GPUs only (integrated graphics ignored but can be manually added by changing GPU column names)
- All feedbacks are welcome especially with unique setups.
- Intel: Core i3/i5/i7/i9 (all generations)
- AMD: Ryzen 3/5/7/9, Threadripper series
- Multi-core: Automatic detection of core count (up to 32+ cores)
"No data found" error:
- Ensure MSI Afterburner logging is enabled
- Check that the log file path is correct
- Verify metrics are being monitored in Afterburner
Missing columns warning:
- Enable desired metrics in MSI Afterburner monitoring tab
- Restart logging to generate complete data
- Check hardware compatibility
Plot style errors:
- Script automatically handles matplotlib/seaborn version differences
- Falls back to default matplotlib style if needed
Memory errors:
- Consider analyzing smaller time windows
- Ensure sufficient system RAM
- Log Duration: 1-4 hours of logging provides good analysis data
- Sampling Rate: 1-second intervals balance detail and file size
- Metrics Selection: Enable only needed metrics to reduce processing time
git clone https://github.com/your-repo/msi-afterburner-analyzer
cd msi-afterburner-analyzer
pip install -r requirements.txt- Follow PEP 8 guidelines
- Use type hints for function parameters
- Include docstrings for all methods
- Add error handling for data processing
- GPU comparison tools
- Real-time monitoring integration
- Advanced statistical analysis
- Custom performance benchmarks
MIT License - see LICENSE file for details.
- MSI Afterburner: For providing comprehensive hardware monitoring
- Pandas/NumPy: For efficient data processing capabilities
- Matplotlib/Seaborn: For powerful visualization tools
- v1.0.0 Initial version has everything that one needs.
- Complete object-oriented rewrite
- Enhanced error handling and logging
- Dual memory analysis (GPU + System RAM)
- CPU power analysis with energy calculations
- Professional visualizations with performance zones