Figure 6 | Scientific Reports

Figure 6

From: Accelerated mapping of electronic density of states patterns of metallic nanoparticles via machine-learning

Figure 6

Computational cost of of PCA-CGCNN methods. Comparison of the computation time for calculations of DOS patterns of metallic NPs via DFT (black) and ML (red). For the DFT calculations, a 2.3 GHz 20 core CPU was used. For the PCA-CGCNN methods, a personal computer with a GTX 2070 GPU and i5-9600 K CPU was used, and the computational times were measured as a sum of training and prediction times.

Back to article page