The olfactory neural network (ONN) is a computational model that was designed based on the structure and mechanisms of the olfactory system. Using real neurophysiological data, the KIII model was developed to match characteristic neural behaviors by iteratively optimizing parameters. This model shares some features with deep learning networks (multiple layers, trainable parameters, nonlinearity), but their implementation involves numerical integration of Ordinary Differential Equations (ODEs). This network has shown a great propensity for pattern recognition/classification problems outside of olfaction. For example, it has demonstrated impressive performance on both a traffic sign and an epileptic EEG dataset. The main contribution of this paper was to increase the computational efficiency of this model so that it can be meaningfully applied to real-world problems. Using the epileptic EEG dataset as a baseline, they found a mean computation time of 26,178 seconds, corresponding to a classification accuracy of 97.87%. In order to deal with this suboptimal latency, they systematically evaluated different ODE solvers and identified the Forward Euler method as the most optimal solver for the olfactory neuron equations. This reduced the computation time to 7,205 seconds and had a 98.19% accuracy on the epileptic dataset. To further improve efficiency, the authors implemented a multithreading-based data parallelism method. They distributed input data across twelve threads while preserving the model’s structural coupling. The final version of their implementation had a computation time of 1,359 seconds. They also verified that model parallelization did not compromise accuracy.