Propagation of spectator particles in IOTA lattice is messed up #310
Conversation
…works) Test of iota propagation without spectator particles (works)
with spectator particles occurs for plain sbends. Refinement of the Booster propagation test showing no issues with spectator particles occur at least over 100 turns. The Booster uses parallel-faced sbends. Cleanup of the iota_propagation test Demonstration that sbend propagation has a problem with spectator particles for both the particle at 0 and the one at the particular value taken from an IOTA simulation.
|
Here's where we are. The problem started when IOTA simulation runs with would lose all particles by 4 turns. Trying to make a small demonstration application I realized that removing the simulation would run if there were no spectator particles. The particle coordinates between the the simulation that runs for 10 turns and the one with spectator particles occured pretty early in the run with very small differences in coordinates that became larger and larger. The first difference occurred after the first sbend in the lattice. We could demonstrate a difference between regular particle propagation and spectator particle propagation with just one plain sbend, but it doesn't occur for the Booster simulation. The difference is that the Booster uses rbends which are converted to parallel-faced sbends. |
|
The two problems:
Are these two different problems or aspects of the same problem? |
Adding tests