2010.06.16 - Frequency Map - Varying fft_turns
For jsh_opt (2010.05.04) sextupole distribution, vary total number of
turns tracked and the number of turns used for FFTs to determine initial
and final tunes.
1024 Turns Tracked
fft_turns = 128
Input file Python script
Frequency map output
fft_turns = 256
Input file Python script
Frequency map output
fft_turns = 512
Input file Python script
Frequency map output
2048 Turns Tracked
fft_turns = 128
Input file Python script
Frequency map output
fft_turns = 512
Input file Python script
Frequency map output
4096 Turns Tracked
fft_turns = 128
Input file Python script
Frequency map output
fft_turns = 512
Input file Python script
Frequency map output
Conclusion
128 turns is insufficient to prevent aliasing effects. There are
differences between 256 and 512 turns, though not nearly as
significant as with 128 turns. The most accurate approach would be 4096
turns of tracking with 512 turns for FFTs. Job time is linear with
number of tracked turns, therefore these jobs take 4x longer than the
1024-turn jobs.
Back