The computational burden for 3D Synthetic Aperture Focusing Technique (SAFT) is large as for each voxel the delay for each acquired A-scan has to be calculated, e.g. O(N5) for N3 voxels and N2 A-scans. For 3D reconstruction of objects which are large in terms of the wavelength, e.g. ≥ (100 )³, the computation of one volume takes several days on a current multicore PC. If the 3D distribution of speed of sound is applied to correct the delays for objects with varying speed of sound the computation time increases further. This overview paper presents the implementations for 3D SAFT developed by the KIT group and discusses their computational performance.