Rothermel_SFIRE model execution time using Anderson 2015 dataset

Workflow information

  • Documentation page:
  • Version: 1.1
  • Date of record creation: 2024-12-03
  • Date of upload to firebench: 2024-12-03
  • Version/tag/commit firebench: 0.3.2a0

Configuration

  • Rate of spread model: Rothermel using firebench.ros_models.Rothermel_SFIRE implementation.
  • Number of point Sobol: 2^10

Specific inputs

  • The environmental variables chosen for this test are:
    • WIND_SPEED from -15 m s-1 to 15 m s-1,
    • SLOPE_ANGLE from -45 deg to 45 deg,
    • FUEL_MOISTURE_CONTENT from 1% to 50%.

Hardware/software description

  • Apple M2, macOS 14.7.1
  • Python 3.10.8

Results

Fig.1 shows the execution time aggregated for all fuel classes (total) and for each fuel class. As fuel category 4 uses a different rate of spread calculation within the SFIRE implementation of Rothermel, a difference in performance is expected. Overall, the performance is very similar for each fuel category and a mean execution time of 6.698 \(\mu\)s over 106,496 samples.

Exec time

Fig. 1 : Execution time boxplot for Rothermel rate of spread model using Anderson13 fuel model. Fliers points not shown on the figure.

Data

  • path to data:
  • 01_generate_data.py: ab95fdc101ae108b3d1592ec0291febf07621e3001c989b0973f2c9d69e48af1
  • 02_plot_data.py: 6e8398298215938a08900555bb04390e4a9281cbb250f6fbc59a66eedb28f854
  • 03_create_record.py: dfb138fe4ccaf999746eefdd9e23cc07453cd5b28ad8247f9eac15d7fa3d84ed
  • firebench.log: eb6ad262459db70de254592248239fe256d4e63f3af69bbbe35ac7d7feaae764
  • output_data.h5: 421036734d13f57d84f9bc0e197255bd70523cf5b91e9e02d320ca805f189c04
  • efficiency_box.png: f1942c22f82c906f84be33c22ff596765e4336727ddfac80c860a470fe37dd7b