Parallel simulation techniques have become a great tool to overcome studies where a large demand of computational effortis needed. One of these cases is the integration of renewable energy in power systems, such as the wind energy. In this paper,it is proposed a highly parallel solver for the simulation of a large number of wind turbines equipped with a doubly-fed induction generator (DFIG), coded using the NVIDIA’s CUDA architecture, based on NVIDIA graphic processing units (GPUs).The implementation involves the solution of a linearized timeinvariant system in state space form, where both 5th and 3rdorder representations of the generator model that is included in a wind turbine are considered. Previously to the CUDA approach, these models have also been implemented in MATLABSimulink and C++ to compare the computational cost. Resultsshow that the CUDA implementation is up to 15 times faster than the other implementations