Fig. 3: Scaling behavior of MALA.
From: Predicting electronic structures at any length scale with machine learning
a Comparison of CPU time between conventional DFT (QuantumESPRESSO) and the MALA framework as a function of the number of atoms. Note that for consistency, slightly different computational parameters have been used for the DFT calculations here compared to the DFT reference calculations in Fig. 4. Additional details on the timings can be found in Table 2. b General workflow of size transferability in MALA.
