The simula-tion of Ground Penetrating Radar (GPR) is one such challenge, where the GPR transducer, subsurface/structure, and targets must all be included in the model, and must all be adequately discretised. The volumetric nature of the FDTD technique means simulations often require extensive computational resources (both processing time and memory). The performance beneﬁts of our GPU-accelerated solver were demonstrated in a GPR environment by running a large-scale, realistic (including dispersive media, rough surface to-pography, and detailed antenna model) simulation of a buried anti-personnel landmine scenario.Ībstract = 'The Finite-Diﬀerence Time-Domain (FDTD) method is a popular numeri-cal modelling technique in computational electromagnetics. We also note that the equivalent Tesla-series P100 GPU – targeted towards data-centre usage – demonstrates signiﬁcant overall performance advantages due to its use of high-bandwidth memory.

We found the cost-performance beneﬁt of the NVIDIA GeForce-series Pascal-based GPUs – targeted towards the gaming market – to be especially notable, potentially allowing many individuals to beneﬁt from this work using commodity work-stations. This is up to 30 times faster than the parallelised (OpenMP) CPU solver can achieve on a commonly-used desktop CPU (Intel Core i7-4790K). Our GPU solver achieved performance throughputs of up to 1194 Mcells/s and 3405 Mcells/s on NVIDIA Kepler and Pascal architectures, respectively. We designed optimal ker-nels for GPU execution using NVIDIA’s CUDA framework. We have developed one of the ﬁrst open source GPU-accelerated FDTD solvers speciﬁcally focussed on modelling GPR. This is exacerbated by an order of magnitude when solving the inverse GPR problem or when using forward models to train machine learning algorithms. Additionally, forward simulations of GPR can necessitate hundreds of models with diﬀerent geometries (A-scans) to be executed.

The Finite-Diﬀerence Time-Domain (FDTD) method is a popular numeri-cal modelling technique in computational electromagnetics.

Open source fdtd