1: \begin{abstract}

2: We present an implementation of the $GW$ approximation for the electronic

3: self-energy within the full-potential linearized augmented-plane-wave

4: (FLAPW) method. The algorithm uses an all-electron mixed product basis

5: for the representation of response matrices and related quantities.

6: This basis is derived from the FLAPW basis and is exact for wave-function

7: products. The correlation part of the self-energy is calculated on

8: the imaginary frequency axis with a subsequent analytic continuation

9: to the real axis. As an alternative we can perform the frequency convolution

10: of the Green function $G$ and the dynamically screened Coulomb interaction

11: $W$ explicitly by a contour integration. The singularity of the bare

12: and screened interaction potentials gives rise to a numerically important

13: self-energy contribution, which we treat analytically to achieve good

14: convergence with respect to the $\mathbf{k}$-point sampling. As numerical

15: realizations of the $GW$ approximation typically suffer from the

16: high computational expense required for the evaluation of the nonlocal

17: and frequency-dependent self-energy, we demonstrate how the algorithm

18: can be made very efficient by exploiting spatial and time-reversal

19: symmetry as well as by applying an optimization of the mixed product

20: basis that retains only the numerically important contributions of

21: the electron-electron interaction. This optimization step reduces

22: the basis size without compromising the accuracy and accelerates the

23: code considerably. Furthermore, we demonstrate that one can employ

24: an extrapolar approximation for high-lying states to reduce the number

25: of empty states that must be taken into account explicitly in the

26: construction of the polarization function and the self-energy. We

27: show convergence tests, CPU timings, and results for prototype semiconductors

28: and insulators as well as ferromagnetic nickel.

29: \end{abstract}

30: