Electric vehicle manufacturers intent on maximizing passenger comfort must consider the acoustics inside the cabin. With conventional engines, the primary noise source comes from the powertrain. For electrical vehicles, road-noise, HVAC noise, and external wind-induced noise become the main critical noise sources manufacturers are looking to reduce. One significant source of internal noise is generated by the side mirror (wing mirror). Design teams can mitigate noise issues by using simulation early in the design process – even before building the first physical prototype. To accurately predict the impact of side mirror noise, high-fidelity computational fluid dynamics (CFD) simulations are required. High-fidelity CFD can be computationally expensive and therefore methods to reduce the computational costs of such simulations are desirable. The reduction in time required to generate meaningful engineering data allows for the opportunity to assess more of the design space and reduce risk in the design process. One such method to achieve this reduction is to utilize advanced High-Performance Computing (HPC) architectures such as General-Purpose Graphical Processing Units (GPGPUs). In this paper, the Simcenter STAR-CCM+ ® solver leveraging the Compute Unified Device Architecture (CUDA) of NVIDIA is used to assess the potential benefits of using GPGPU architecture. DES simulations using the Perturbed Convective Wave aeroacoustics model are applied to an example side-mirror configuration, the AeroSUV model. Comparisons between numerical simulations using GPGPU and traditional Central Processing Units (CPUs) are undertaken. The accuracy and computational cost of the numerical simulations is discussed. CPU and GPU configurations will be provided with the extended abstract. Frequency spectra at receivers located both on the side glass of the car and in the air volume surrounding the wing mirror are compared, while monitoring simulation turnaround time on CPU and GPGPU. It is demonstrated that the use of GPGPU architectures can result in high-fidelity aeroacoustics modelling within the required turn-around times of engineering design cycles at significantly beneficial computational cost.