Table 2 Computation efficiency of training and predicting processes of the numerical model and P2M

From: A machine learning-based prediction-to-map framework for rapid and accurate spatial flood prediction

  

Numerical model

P2M

Training

Number of cores

240

Run time

720 h

Predicting

Number of cores

480

1

Run time

4 min for initialization + 2 min / simulation hour

4 s

  1. All training and forecasting processes were implemented on CPUs on LSU’s QB3 cluster (https://www.hpc.lsu.edu/resources/hpc/system.php?system=QB3).