The new EMPIRE XPU solver shows a re-organized new simulation tab which gives a comprehensive status on all currently running simulation activities. Operation buttons, job trees, log and convergence windows have been introduced.

During simulation EMPIRE XPU continuously evaluates energy and field monitor convergence and terminates the job if accuracy levels have been obtained. The post-processing is started subsequently for each simulation job.

Single and Cluster Simulation

EMPIRE XPU features the fastest FDTD solver on market. Its unique XPU algorithm is a software driven acceleration which runs multiple time steps in the last level caches before exchanging and reassemble data with the main memory.

  • Powerful XPU technology.
  • Parallel simulation is supported on multiple cores and CPUs
  • Remote simulation capabilities
    • Job queuing, scheduling and result retrieval
    • Sophisticated remote host management showing status, user and available performance
  • Cloud simulation support (Amazon AWS and Microsoft Azure)
  • Cluster Solver
    • Solver for extremely large jobs to run efficiently on multiple PCs.
  • Multi port processing
    • Creation of full admittance matrix Y, impedance matrix Z and scattering matrix S
    • Automatic touchstone file generation available
  • Thermal Solver
    • The temperature distribution can be obtained with a thermal simulation based on thermal sources, thermal parameters of objects, heat sinks and thermal simulation parameters.
    • Electromagnetic losses can be used as thermal source
    • Convection and radiation heat sink coefficients on surfaces

Variable Sweeps and Optimization

Geometrical and physical properties can be parameterized and automatically varied for tolerance, yield analysis or optimization. If available, jobs can be distributed on remote hosts in a network. Each sweep or optimization step creates unique folders for tracking the result.

  • Variable range, lists or equations
  • Complete result folders for each variable combination
  • Collection of different optimizer (Gradient, Simplex, Newton, Derivative-free, Random, …)
    • Range (e.g. frequencies, angles) or single value goals (e.g. efficiency)
    • Multiple weighted goals
  • Distributed or sequential execution

Far Field Processing

Far fields are obtained by recording the near field on a Huygens surface (simulation boundary) and applying a far field transformation during post-processing. This allows a compact simulation domain which is only a little larger than the radiation source itself.

  • Multi-frequency or broadband recording
  • Arbitrary far field cut angles or 3D radiation patterns
  • Directivity, Gain, Total gain or electric field normalization
  • MIMO envelope correlation coefficient calculation
  • RCS (mono-static) or bi-static scattering cross section analysis

Near Field Processing & SAR

For analyzing fields in human bodies several averaging methods are defined in compliance standards. A low frequency algorithm can be applied in case of very low frequencies, e.g. coil charging

  • Specific Absorption Rate (SAR) with mass averaging methods
  • Averaged Current Density (ACD) with area averaging
  • Averaged internal electric field (EIAV) with volume averaging
  • Averaged power density (PD) with area averaging