Resolve "Fix tests that use default `P0` and `FROMFILE`"

Closes #109 (closed)

added Cleanup label

added 1 commit

• f3b68d9a - Correct BEAM momentum to agree with FROMFILE distribution.

Compare with previous version

marked this merge request as ready

requested review from @ext-calvo_p

Kick-Test-1 still fails, but I can't catch the error

Error: size of stat variables (rms_x) dont agree!
       size reference: 142, size simulation: 143

It seems a precision bug, because the time of the first step prints in the stdout differs from reference

Yeah I also got this and I don't understand. I checked with OPAL master and it also fails, so it's not a bug that we introduced now.

I was wrong, sorry. Actually with the master branch it fails only if I give it P0, but if I don't, then it uses the default P0 and it doesn't fail.

The statistics in the stat file are written every ten steps, but if we give the correct P0 the chosen timestep is a bit smaller so it writes to the stat file 143 times instead of 142, that is why it is failing.

IMO this was a bug before this MR, because the timestep was chosen according to the momentum in BEAM instead of being chosen according to the momentum in the particle file.

Yes, you're right. So I would update the reference results with the correct input momentum.

I don't understand why the differences are so big. The only difference is that I am giving the correct P0.

This is because the energy of the reference particle was very different from the energy of the beam. It’s the energy of the reference particle that determines the field strength of the kick (corrector) but it’s path isn’t influenced by the corrector.

added 1 commit

• 7bfffde6 - Update Kick-Test-1 test with correct P0.

Compare with previous version

added 1 commit

• 37ccd73f - Remove unnecessary comment line.

Compare with previous version

approved this merge request

added 1 commit

• 45c3ec8f - Remove unnecessary definitions of energy in AWAGun-TrackBack-1 test.

Compare with previous version

merged

mentioned in commit ec6d2583

Sorry, I clicked the wrong button. Wanted to approve this MR but merged it instead.