The efficient use of energy for heating and cooling of indoor environments requires an accurate prediction and analysis of radiative heat transfer. Therefore it is necessary to use modern computer methods as otherwise the computational costs may become higher than for convective heat transfer, which is known to be a huge computational problem. A key problem in calculations of radiative heat transfer is the problem of mutual visibility arising in the determination of view factors. This is the same challenge that global illumination has to cope with which is one of the fundamental topics in computational graphics. The visibility problem is efficiently solved by modern graphics hardware. Therefore an OpenGL-based algorithm is developed to quickly and accurately calculate view factors for arbitrary, complex geometries. Theoretical and implementation details of the applied methods are given. We demonstrate the advantages of the developed computational method by a virtual test room for a tubular radiator, a heating of a warehouse by ceramic infrared heaters and the heat transfer in a car cabin.