13.6 Highlights of What’s New in Version 2019.3.0

Here is a summary of what’s changed since release 2019.2.1. Historical change logs are collected in Section 13.

13.6.1 Bug Fixes for Elements

• The matrix for the SBEN element was incorrect when the length was negative (which is needed for back-tracking). A. Zholents (ANL) and Y. Park (UCLA) helped identify the problem.
• The integrator used for the CSBEND element with the expanded Hamiltonian (EXPAND_HAMILTONIAN=1) was very inaccurate and yieled poor results unless N_KICKS was large. This problem, pointed out by Z. Duan (IHEP), was fixed.
• The interpolation used for UKICKMAP would previously produce invalid values for particles near the upper (y > 0) and left (x > 0) edges of the grid. In essence, the interpolation assumed the kickmap was periodic in x and y.

13.6.2 Bug Fixes for Commands

• DA trimming (to make the xClipped and yClipped columns, as well as the Area parameter) was not working properly in the parallel version for full-plane runs. This was fixed. In addition, the algorithm was improved for both the serial and parallel versions to iterate the trimming until it converges.
• The multi-gaussian and multi-lorentzian feature of the ion_effects command, added in the last release, proved unreliable, as reported by B. Podobedov (BNL). Several improvements and bug fixes were implemented that should improve matters.

13.6.3 New and Modified Elements

• The SHRFDF element was added, which models a deflecting rf cavity using a space harmonic expansion. This was implemented by Y.P. Sun (APS). See [55] for details.
• The LSCDRIFT element, which models longitudinal space charge, can now have its effective length set automatically to correspond to the length of the upstream element.

13.6.4 New and Modified Commands

• The run_setup command has a new parameter, back_tracking, which allows invoking a limited back-tracking capability. See the entry for run_setup for more details. This is an experimental feature and users are encouraged to report problems to the forum.
• The ion_effects command was improved in several ways:
  • The default distribution fitting parameters were modified to give improved convergence
  • The default distribution fitting criterion is now the sum of the maximum fractional absolute deviation over the histogram and the absolute fractional deviation of the ion charge. This makes it less likely that overfitting will result in large spikes in the distribution.
  • The new ion_histogram_max_bins parameter allows restricting the maximum number of bins.
  • The new ion_histogram_min_per_bin parameter allows setting a requirement on the minimum number of macro ions per bin.
  • The new freeze_ions_until_pass and freeze_electrons_until_pass parameters allow “freezing” the motion of the ions and electrons until a specified pass number. This is useful for diagnostic purposes.
  • The new pressure_factor parameter allows multiplying all the pressure profiles by a common factor.
• The matrix_output command can now print the full matrix in a form accepted by Mathematica.

13.6.5 Changes Specific to the MPI Parallel Version

• Some apparent MPI-related issues were resolved for the ion_effects command.

13.6.6 Changes Specific to the GPU Version

The GPU version continues to be an alpha release and contains bugs. Users are encouraged to check results against the serial or parallel versions and report issues to the developers.

• None.

13.6.7 Changes to Related Programs and Files

• The touschekLifetime program now reports the value of deltaLimit in the output file, whether that value is given explicitly or computed via the rf voltage.
• Added the program sdds5x5sigmaproc, which computes the 5x5 sigma matrix (i.e., all elements except those related to the time coordinates) from a quadrupole scan.