GPT/Pulsar Logo

Pulsar Physics

and the General Particle Tracer (GPT) code

Run button Introduction: Over the past years the General Particle Tracer (GPT) package has become a well established simulation tool for the design of accelerators and beam lines. GPT is based on full 3D particle tracking techniques, providing a solid basis for the study of 3D and non-linear effects of charged particles dynamics in electromagnetic fields. All built-in beam line components and external 2D/3D field-maps can be arbitrarily positioned and oriented to simulate a complicated setup-up and study the effects of misalignments. An embedded fifth order Runge-Kutta driver with adaptive stepsize control ensures accuracy while computation time is kept to a minimum. GPT provides various 2D and 3D space-charge models, including a sophisticated 3D particle-mesh method that scales O(N) in terms of CPU time. Because of its modern implementation, GPT can be easily extended by the user to perform highly specialized calculations for specific applications. Hierarchical data analysis, automatic parameter scans, built-in optimization and graphical output allow for fast and detailed interpretation of the simulation results.


Highlights of GPT version 3.5: The latest GPT version 3.5 release contains a numer of performance improvements and the following new features:
Spin tracking (fully relativistic)
Memory mapped 3D fieldmaps, see below.
Support for non-uniform meshline distributions in 3D fieldmaps.
MacOSX release with binaries for both Intel and Apple sillicon (M1) processors.
Coherent Synchrotron Radiation (CSR) module standard included (used to be a seperate product.
Simplified notation for 3D coordinate transforms.
Improved accuracy for 1D field-maps when derivative information is present.

Note: Memory mapping ensures that a field-map is shared between GPT processes. This signifiantly reduces memory pressure during multi-core optimisations or when the same fieldmap is used many times within the same GPT simulation.

GPT-BEM (Boundary Element Method): Since GPT version 3.3 there is a seperate package that calculates electrostatic and magnetostatic fields in full 3D with a precision that is unreachable by normal Finite Element codes. This unmatched precision is necessary for the design of the next generation electron microscopes and related applications. Since this is a very specialised field, our BEM solver is sold as a separate product.
Hierarchical Boundary Element Method solver
Support for millions of surface triangles
Achieved relative accuracy >10-6
Electrostatic and magnetostatic
Aberration analysis up to 5th order
Multipole output up to 7th order


Further information: The GPT Versions page can be used to view additional technical information per major release and platform availability. Refereed publications about GPT and GPT related projects can be found on our Publications page. Newly developed beamline components, updates of the User Interface and an up-to-date bug list can be found on our News page. A description of our own research is listed on our Projects page.


Is GPT an appropriate tool for you? Although the capabilities and graphical user-interface of GPT dramatically simplify the task of accelerator and beamline design, the package is not for the novice. The learning curve is fairly steep, simply because using a particle tracking code requires professional skills regardless of the tools.


 GPT plotting Field import Collector design Root-finding and solving GPTwin User Interface


Documentation and development: GPT is not a black-box simulation package: All internal calculations and all 3D electromagnetic field configurations are fully documented. Furthermore, the source code of all built-in elements is available, allowing users to easily develop models for custom beam line components. GPT has been under constant development since its first release in 1996 as a result of the joined effort of Pulsar Physics and all GPT users. We feel very privileged to be able to continue this path.

Dr. S.B. van der Geer, Dr. M.J. de Loos
The authors of the GPT code.