calculations for molecules with benchmark accuracy

In traditional GW implementations, the computational cost is growing as O(N⁴) in the system size N, which prohibits their application to many systems of interest. I present a GW algorithm in a Gaussian-type basis, whose computational cost scales with N² to N³. It will be shown that large minimax grids and resolution of the identity with the truncated Coulomb metric improve the accuracy of the low-scaling GW algorithm to < 0.01 eV for the GW100 test set. Large-scale applications of low-scaling GW will be discussed.