Framework

Sim4Life integrates Python, a powerful scientific scripting language with access to a vast range of third-party libraries. The Sim4Life Python application programming interface (API) enables users to parameterize and automate tasks such as geometric modeling, simulation setup, and postprocessing. It also allows for the development of custom tools and standalone applications.

Key Features

• Scripting access to all Sim4Life features at every level (modeling, grid, simulation, postprocessing, etc.)
• Includes the Enthought Python package, offering powerful libraries such as SciPy, NumPy, and pandas
• Possibility to use the entire Sim4Life API programmatically without launching the GUI
• Fully interfaced scripting capabilities for automation, batching, parametrization
• Integrated Scripter for writing, running, and debugging scripts inside Sim4Life
• "To Python" export turns models, simulation setups, and analyses built in the GUI into reusable scripts
• Consistent, fully documented Python API with a searchable API reference
• AI Assistant generates Python scripts from plain-language requests and explains each step (see the AI Assistant tab)
• JupyterLab with full access to the Sim4Life Python API in Sim4Life.web
• Enhanced tabbed script editor allowing multi-file selection and opening
• Support for writing, executing, and debugging scripts for Sim4Life via Visual Studio Code
• Unlimited user customization (create custom tool libraries and build a personalized Sim4Life environment)
• Launch batch simulations, parametrization, and optimization tasks
• Automated extraction of various parameters and result data

Sim4Life provides high-performance computing (HPC) capabilities to support the investigation of complex and realistic models. Multi-threaded execution for modeling, meshing, voxeling, and postprocessing enables parallel processing of demanding tasks without disrupting the workflow. A fully integrated centralized task manager (ARES) efficiently manages all computationally intensive tasks on local machines or in the cloud. Two sets of CUDA-based libraries for graphics processing unit (GPU) acceleration are available for EM-FDTD: Acceleware and ZMT CUDA.

HPC

• Acceleware libraries for EM-FDTD
• ZMT CUDA libraries for EM-FDTD, ACOUSTICS, and THERMAL
• Support for thin sheets, dispersive media, and lossy metals (Acceleware)
• Option to add support for Subgrid (local refinement) with Acceleware
• Multi-GPU support
• Improved OpenMP parallelization
• Servers/Workstations, single- or multiple GPU systems
• MPI engine for Electromagnetic Quasi-Static (EM-QS) solvers

ARES (Desktop)

• Fully integrated centralized task manager
• All functionality, including remote computing and HPC, is seamlessly integrated into the Sim4Life and Python frameworks
• Parallel processing of computationally intensive tasks (e.g., meshing, simulation, postprocessing)
• Remote execution via sim4life.web (e.g., AWS), localhost, GPU server, MPI , p2p, etc.
• Heavily multi-threaded execution for modeling, meshing, voxeling, and postprocessing
• Queuing control instances, statistics
• Solver control via the Sim4Life GUI
• Intelligent job submission

Postprocessing in Sim4Life provides comprehensive insight into intricate simulation results and measurement and imaging data through advanced visualization and analysis capabilities. Specialized viewers for medtech simulations (e.g., volume rendering, streamlines, maximum intensity projection, surface fields on arbitrary 3D structures) and calculators for recurring tasks (e.g., resampling, filtering, functional evaluation) on both unstructured and structured field data are included

Key Features

• Novel Visualization Toolkit (VTK)-based pipeline architecture, template pipelines (on-demand driven work/data flow)
• Support for unstructured and regular field data
• A variety of processing algorithms (e.g., filtering, evaluation, etc.)
• Volume rendering on GPU, streamlines, maximum intensity projection, interpolation on arbitrary 3D structures, surface field rendering, and more
• Calculators, processing algorithms such as resampling, filtering, evaluation, cropping
• Computation of power balance, including losses in SiBC
• Functionality to compress 3D field data for output files of FDTD simulations, reducing disk space consumption
• 2D XY plots, polar plots, Smith chart, and animations
• Radiation pattern (3D in model, 2D XY/polar)
• EM fields (Avg P, B, D, E, H, J, S, Energy, etc.)
• Radiation and far-field data, efficiency, TIS, TRP, Radar Cross-Section
• SAR/absorption (Av. 1g/10g/arbitrary IEEE1529, distributions, dP/dV)
• Full multi-port S-Parameter extraction
• Solid/material selective extraction/processing
• User-friendly line field extraction tool
• Statistics processing, visualization, and extraction into tables
• Sleek option to slice a model from the slice field viewer and generate cross-section visualization of the (selected) model parts
• Total current flux normalization option for low-frequency simulations driven by a potential difference
• 3D viewers (2D planar slice, overlay on surfaces conformal/Yee, vectors/arrows, ISO surfaces, etc.)
• Novel 3D vector and streamline viewers
• Combination/comparison and visualization of multiple results
• Overlaid visualization of model/voxels/results
• Full support for Python scripting, enabling access to all postprocessing features
• Data processing in both time and frequency domains
• Data export into third-party editors
• Viewer exports to image data
• Automatic Cache Save/Load extracted results

The Sim4Life AI Assistant is available throughout the application and combines product knowledge with support for Python scripting, CAD modeling, simulation setup and review, and postprocessing. With the new “Ask Sim4Life” keyword, users can adapt parameters, models, simulations, and grids through plain-language instructions that take the current project context into account. The AI Assistant works in the same interface on the desktop application and the web platform. In a later version, users will be able to bring their own tokens for their preferred model provider.

Text-to-Simulation in practice: In a peripheral nerve stimulation study, a single request to the AI Assistant builds a 3D nerve from a labeled 2D cross-section and fits cuff electrodes to its surface. iSeg delineates the fascicles, and neuron models are seeded within each fascicle and routed along realistic 3D trajectories, producing a simulation-ready nerve. Mesh refinement and convergence analysis then follow best practices set by our physics and biomedical experts. The results show action potentials propagating along the nerve fibers.
 

Key Features

• AI Assistant available throughout the application
• Built-in Sim4Life product knowledge
• Support for Python scripting, CAD modeling, simulation setup and review, and postprocessing
• “Ask Sim4Life” keyword for context-aware, plain-language adaptation of parameters, models, simulations, and grids
• Text-to-Simulation: complete workflows from model creation to results, built from plain-language requests
• Mesh refinement and convergence analysis following expert-configured best practices
• Same interface on the desktop application and the web platform
• Bring-your-own-token support for preferred model providers (in a later version)