Benchmarking of a TLM solver on mobile devices sharing a common hardware architecture

Modern mobile communication devices share a common set of hardware and software features. Within a brand, it is also common for similar hardware architectures to be used, and to be paired with the same operating system. Whilst mathematically trivial in nature, the Transmission Line Matrix (TLM) method is computationally-intensive, and therefore varies in performance dependant on available resources. This paper investigates how the performance of a mobile Computational Electromagnetics (mCEM) laboratory iPhone application is effected when executed on devices with different hardware configurations, but that are available from the same manufacturer. Hardware differences across the chosen platform are first discussed, with performance predictions made based on the device specifications. TLM is suitable for parallel computing implementations, and therefore potential multi-threading optimisation strategies are discussed for applicable devices. Results from the application performance benchmarking are presented, and demonstrated that the iPad 2 offered the best processing performance of the devices tested.