OptiFDTD provides two types of FDTD simulations:
- 32-bit simulations (performed by 32-bit Simulators).
- 64-bit simulations (performed by 64-bit Simulators).
These simulation options differ in certain aspects of supporting features or.
The most important difference between these two simulator types is the memory limitation imposed on 32-bit simulators. All the 32-bit applications may utilize up to 2GB-4GB of memory for its purposes, dependent on the operating system.
In case of 64-bit applications (simulators), this limit is in the range of Tera-Bytes. In the current Windows 64-bit this limit is around 16TB, which is far beyond RAM capacity supported by standard PC configurations.
The 32-bit simulations can be executed under Windows 32-bit and Windows 64-bit operating systems. In case of Windows 32-bit, the operating system can provide to the simulator up to 2GB-3GB of memory, while the remaining 2GB-1GB is used by the operating system. Under Windows 64-bit, the operating system can provide the 32-bit simulator with up to 4GB of memory, while it will utilize another ~4GB for its operating system memory mapping needs.
In case of 64-bit simulators, they can be executed only under Windows 64-bit operating system. One cannot execute a 64-bit application under a 32-bit operating system.
All the references to 64-bit Windows operating systems assume usage of a processor supporting EM64T architecture. In layman terms it means “hardware support for 32- bit and 64-bit applications”. The currently manufactured Intel or AMD processors support this architecture, for example Intel’s Core 2 based multi-core processors.
In short – the main difference between the 32-bit and 64-bit simulations is that 64-bit simulators are capable of processing much larger designs. Due to the amount of data volume processed and generated by 64-bit simulators, the approach to the simulation result handling has changed.
The notes below describe the main differences.
— Layout size
- 32Bit simulation can only use maximum 2GB of RAM. This means for a 3D simulation, the maximum mesh point number that can be handled will 32Bit simulation can only use maximum 2GB of RAM. This means for a 3D simulation, the maximum mesh point number that can be handled will be roughly 300X300X300. When the mesh size exceeds this value, 64bit simulation is recommended.
— Time domain dynamic response
- 32bit simulation shows the dynamic field response in Simulator window
- 64bit simulation does not show the dynamic field response. Simulation process (percentage of the simulation) is displayed. The snapshot for time domain response in the certain time step for user selected observation area/line/point can be saved and viewed by Optiwave graph viewer
— Volume DFT field results (steady state response)
- 32bit simulators calculate and record volume DFT field response (steady state response) for center wavelength
- 64bit simulators do not store the DFT field response for the whole volume anymore. However, the steady state response is calculated and stored for Observation Area/Line/Point for user-defined wavelength range. This data can be observed in the Analyzer, after completion of the simulations.
— Observation Detector and Spectrum Analysis
- 32bit simulations will record all the time domain response in the Observation detector (observation area, line). User can specify any wavelength to perform the DFT spectrum analysis.
- 64bit simulations perform the spectrum analysis for observation line and observation area in simulator. No time domain response is recorded for Observation line and area. Spectrum results, for user pre-defined wavelength, are saved to analyzer (result) file. In the Analyzer, the user can select a wavelength from the previously defined list and observe the corresponding spectral results in observation area analysis. The user may define a wavelength range in the Layout Designer, in the “64-bit Simulation Parameters” dialog box (click the Spectrum button).
- One must set an observation detector (Observation Point, Observation Line, and Observation Area) to collect calculation results for 64bit simulation.
- 32bit FDTD simulation supports single processor.
- 2D 64bit FDTD simulation support single processor in OptiFDTD 8.0.
- 3D 64bit FDTD simulation support multi-processor (core)
- 2D TFSF excitation is only supported in 2D 32bit simulation in OptiFDTD 8.0
— Tutorial lessons
- Tutorial lessons specify the 32bit simulation to demonstrate the product features. However, one can always use the 64-bit simulations to perform calculations, or simulate an equivalent example configured for 64-bit simulations. Typically, the 64bit simulations are faster.
- If you use a 64-bit simulation instead of 32-bit one, then make sure that the following items are addressed:
— Please define an Observation Area/Line to collect steady state response, transmission function and reflection function
— If the input wave is Gaussian Modulated Continuous Wave, please define the wavelength range (click on the “Spectrum” button) in the “64-bit Simulation Parameters” dialog box.
This chapter is to help you familiarize yourself with the OptiFDTD Designer
environment. Please use the following lessons for more detailed simulations on each