Faster and more accurate stochastic GW*

I will present recent developments in stochastic approach to the GW approximation, which further accelerate the calculation of quasiparticle energies and increase their accuracy. A new concept of sparse stochastic compression is used to speed up stochastic approaches and leads to an overall decrease of statistical errors in large finite and periodic systems. Computation of quasiparticle energies and gaps for systems with up to N_e >10,000 electrons is thus feasible with only small statistical fluctuation (± 0.05 eV) and consuming < 2000 core CPU hours. Further, I will present an efficient scissors-like GW self-consistency approach that can be implemented at zero additional cost. This result is a simple modification of the time-dependent G0W0 and enables an a posteriori self-consistency cycle applicable to large systems.