GPT Versions: All recent GPT versions run under Microsoft-Windows, Linux and Mac OS X. However, the Linux and Mac releases do not contain the graphical 'GPTwin' user interface, and are therefore not able to produce graphical output. The main feature of GPT version 3.1 is a different tracking engine with significantly faster and more accurate output. Our multi-objective genetic optimizer was first introduced in GPT version 3.2 and the main addition to the latest 3.3 release are wakefields and CSR capabilities.
More detailed information about the different GPT versions is shown in the GPT Versions and the GPT platforms tables shown below. See also our Features page for more information about the capabilities of GPT.
|GPT core technology|
|- 3D relativistic equations of motion solver|
|- Fast and accurate time/position output|
|- Wakefields and CSR1)|
|- Multigrid Particle-In-Cell|
|- Granularity/stochastic effects|
|Beam line modeling|
|- 3D positioning of all beamline components|
|- 1D/2D/3D field-maps|
|- Custom elements|
|- Collector design and boundary elements|
|- FEL interaction|
|Pre- and postprocessing|
|- Multi-dimensional parameter scans|
|- Hierarchical data analysis|
|- Multi-objective genetic optimizer|
|Platforms for GPT version 3.3||MS-Windows||Mac OSX||Linux 4)||MPI-cluster4)|
|'GPTwin' Graphical User Interface 2)|
|Multi-core parameter scans|
|Muti-core multi-objective optimization|
|64-bit command-line executables|
|MPI parameter scans|
|MPI multi-objective optimization|
1) The Coherent Synchortron Radiation (CSR) module is very CPU intensive and only present in the MPI-Linux version.
2) The Linux and Mac versions of GPT do not ship with a graphical user interface. We therefore recommend converting GPT and GDFA output to ascii or SDDS, and subsequently using public domain plotting packages such as GNUplot and the SDDStoolkit. Alternatively, scripts are available on the GPT Community page that convert binary output to Matlab and HDF formats. On-line help is always available as PDF manuals.
3) Running the MPI-cluster version requires one 'per machine' license for that machine in addition to an individual 'per user' GPT 3.2 license for each active GPT user.
4) Intel based SUSE or RHEL Linux. Other systems on request.