13.27 Highlights of What’s New in Version 33.0, March 3, 2017

Here is a summary of what’s changed since release 32.0. Historical change logs are appended to the end of this manual.

This version includes an alpha release of GPU-enabled code. The original GPU code was developed by Tech-X corporation [51], with further work by R. Soliday (APS).

13.27.1 Bug Fixes for Elements

• The SREFFECTS element now correctly computes the equilibrium horizontal and vertical emittances when J_x≠1. Previously, the computation used an equation that implicitly assumes J_x = 1.
• The MALIGN element could cause spurious integer changes in the reported tunes if the DZ parameter was negative. This problem, reported by V. Sajaev (APS), was fixed.
• A memory management bug related to the systematic and random multipole data store was fixed. This in principle affected KQUAD, KSEXT, and other elements using the SYSTEMATIC_MULTIPOLES and RANDOM_MULTIPOLES features. In testing, no effect was in fact observed.

13.27.2 Bug Fixes for Commands

• The correction_matrix_output command command were ignoring the monitor calibrations (MONI, HMON, and VMON) values when use_response_from_computed_orbits = 1. This was reported by V. Sajaev (APS).
• The steering_element command no longer aborts even if the declared steering corrector appears not to kick the beam. This allows using unusual controls such as path length to steer the beam. This issue was pointed out by V. Sajaev (APS).
• The load_parameters and save_lattice commands incorrectedly saved the edge angles and other edge-related quantities for bending magnets that were reflected. This issue was fixed. Previously-saved parameter files should be modified (e.g., remove the edge parameters) unless the magnets had the same parameters for the entrance and exit. This problem was reported by Y. Li (BNL).
• The rf_setup and moments_output commands will now run in a loop with find_aperture, momentum_aperture, and frequency_map operations, if set for per-step execution. Previously, this would only happen for the track, analyze_map, and touschek_scatter commands.

13.27.3 New and Modified Elements

• The EKICK, EHKICK, and EVKICK elements now support inclusion of multipole errors linked to the correction strength.
• The steering kicks and steering multipoles in the KQUAD element are now implemented in the body of the element, rather than at the ends.
• The WATCH element was improved so that the dt column in coordinate-logging mode and the dCt column in parameter- and centroid-logging modes are more useful. In particular, in normal cases these will now more reliably be centered on zero. One can also provide a reference frequency relative to which the reference time is defined. This improvement grew out of discussions with J. Calvey and T. Berenc (APS).
• The reported phases of the beam- and generator-induced parts of the voltage for the RFMODE element RECORD file are now computed using a method that should be more reliable. This improvement grew out of discussions with J. Calvey and T. Berenc (APS).
• The RECORD output from the RFMODE element now includes the phase of the net cavity voltage. This was requested by M. Venturini (LBNL).
• The RFMODE element now supports injection of noise into the rf source and low-level rf system. This is based on discussions with T. Berenc (APS).
• The SCRIPT element can now import particleID data from the script without attempting to use this information for lost-particle accounting. This provides better functionality when the particleID is used for other purposes, such as bunch membership.
• The TFBPICKUP and TFBDRIVER elements, used for bunch-by-bunch feedback, now allow 30-term FIR filters, up from 15 turns in earlier versions.
• The TFBDRIVER element now accepts specification of the frequency and phase of the driver cavity.
• Aperture enforcement inside KQUAD, KSEXT, KOCT, KQUSE, CSBEND, and CSRCSBEND elements has been improved. In particular, the ELLIPTICAL, EXPONENT, YEXPONENT, and OPEN_SIDE parameters of MAXAMP are now implemented. In addition, for the fourth-order integrator, the apertures are no longer asserted at each integration step, but only after each slice (or “kick”, to use the misleading terminology of the element parameters).
• Added the ALLOW_LONG_BEAM parameter to the ZLONGIT and ZTRANSVERSE elements.

13.27.4 New and Modified Commands

• The bunched_beam command can now be set to take the fully-coupled 6D bunch parameters from the calculations of the moments_output command, provided the latter is used to compute matched, equilibrium parameters. This was requested by forum user duanz.
• Added occurrence and positional filters for the steering_element command. This was requested by V. Sajaev (ANL).
• Several informational printouts for the touschekScatter command are no longer shown by default, but only if the verbosity control is set to a non-zero value. This makes short runs more efficient.
• Compared to previous versions, the lost-particle data file (losses file requested by the run_setup command) will exhibit changes in the order in which particles are recorded. This was a result of reworking the code for lost particle management.

13.27.5 Changes Specific to Parallel Version

• None.

13.27.6 Changes to Related Programs and Files

• The program madto was renamed elegantto, to more accurately reflect what it does. It will now translate elegant lattice files into MAD8 format.