Graphical Processing Units (GPUs) have evolved at a large pace, maintaining a processing power orders of magnitude higher than Central Processing Units (CPUs). As a result, the interest of using the General-Purpose computing on Graphics Processing Units (GPGPU) paradigm has grown. Nowadays, effort is being put to study probabilistic search algorithms like the Randomized Search Algorithms (RSA) family, which have good time complexity, and thus can be adapted to massive search spaces. One of those algorithms is Rapidly-Exploring Random Tree (RRT) which reveals good results when applied to high dimensional dynamical search spaces. This work consists in the design, exploration and study of the use of GPGPU-based parallelization techniques in order to improve the application of RRT to racing videogames. To approach such study, a new variant of the RRT algorithm called Iterative Parallel Sampling RRT (IPS-RRT) was developed and a bot for the TORCS open source racing game was built. The results show that, although accesses to the GPU’s memory present high latency, the use of GPGPU-based techniques like the one of this work can still improve not only the planning computational efficiency but also the quality of the returned solutions, as GPU IPS-RRT achieved temporal improvements in big problem sizes (when generating 6400 states) and lap time reductions of around 19%.