RestartTest-2 broken

This test ran yesterday, it failed but wasn't broken. However I don't see an apparent cause in the committed changes.

It's my mistake. I get following error

Error{0}> 
Error{0}> *** User error detected by function "DataSink::writeGridLBalHeader()"
Error{0}>     Can not write grid load balancing for non-AMR runs.
Error{0}>     Can not write grid load balancing for non-AMR runs.

I'll fix it shortly.

@kraus Above is a mistake but is OPAL anyway compiled with ENABLE_AMR? Do you know?

The test runs on my system without any issue (after fix src@93eede1a since I enabled AMR). BTW, I just realised that other tests would have been broken as well if ENABLE_AMR=1. It's too early ...

As far as I know OPAL isn't compiled with AMR. Other tests are possibly broken as well. But the script that runs the regression tests doesn't delete old results (this is fixed in the repo, @gsell should check out the latest changes). The consequence is that tests seem to run smoothly although they're broken.

What I see in the stat-file is that the last 7 columns aren't written in the restart.

These columns are probably the number of particles that are stopped in collimators (?).

The last column is partsOutside. The description tells outside n*sigma of the beam.

The reference data is anyway in the old format, i.e. there are columns missing like

num	name
18	ref_x
19	ref_y
20	ref_z
21	ref_px
22	ref_py
23	ref_pz
41	dt

According to http://amas.web.psi.ch/opal/regressionTests/master/results_2018-02-01.xml the test is not broken, but still failing.

@kraus May I close this issue?

Achim re-branding free tests... Did you change something in the mean time?

closed

Auto-completion... it should have been re-ran the tests...

No, I did not update the reference solution (if you mean that). But I can generate a new stat file for this test.

The only thing I did so far, was a fix of the AMR problem, but this was no issue here.

In order to generate a new reference, I ran 3 turns and compared my result to Restart-2.stat.bak. I see discrepancies, but I remember that the LF-2 algorithm had a wrong bunch position and got fixed as mentioned in src#124 (closed).

@kraus If you agree, I update the reference with this.

I meant changes in the code.

Just go ahead.

I struggle a bit with this test. When I create the reference data I have 1000 particles at the end

OPAL{0}> *** Finished turn 3, Total number of live particles: 1000
OPAL{0}> 
OPAL{0}> * ---------------------------- DONE TRACKING PARTICLES -------------------------------- * 
OPAL{0}> *
OPAL{0}> * Finished during turn 4 (3 turns completed)
OPAL{0}> * Cave: Turn number is not correct for restart mode
OPAL{0}> 
OPAL{0}> * *********************** Bunch information in global frame: ***********************
OPAL{0}> * ************** B U N C H ********************************************************* 
OPAL{0}> * NP              = 1000
OPAL{0}> * Qtot            =       19.743 [pC]         Qi    =       19.743 [fC]
OPAL{0}> * Ekin            =      79.076 [MeV]         dEkin =      21.205 [keV]
OPAL{0}> * rmax            = (      2.22807 ,      0.01805 ,      0.00111 ) [mm]
OPAL{0}> * rmin            = (      2.20803 ,     -0.00856 ,     -0.00108 ) [mm]
OPAL{0}> * rms beam size   = (      3.97397 ,      3.85552 ,      0.46607 ) [mm]
OPAL{0}> * rms momenta     = (  1.29972e-03 ,  5.84683e-05 ,  2.96932e-04 ) [beta gamma]
OPAL{0}> * mean position   = (      2.21793 ,      0.00625 ,     -0.00000 ) [m]
OPAL{0}> * mean momenta    = ( -6.48693e-04 ,  4.19114e-01 , -2.05116e-06 ) [beta gamma]
OPAL{0}> * rms emittance   = (  5.91650e-06 ,  4.94873e-07 ,  2.14364e-07 ) (not normalized)
OPAL{0}> * rms correlation = (  8.75455e-01 , -3.76965e-01 , -7.56894e-01 )
OPAL{0}> * hr              = (      3.01048 ,      3.81883 ,      0.33937 ) [mm]
OPAL{0}> * dh              =   2.00000e+00 [%]
OPAL{0}> * t               =      355.380 [ns]         dT    =       61.698 [ps]
OPAL{0}> * spos            =        40.379 [m]
OPAL{0}> * **********************************************************************************

But when I restart it only finds 999 particles

Restart from hdf5 format file RestartTest-2.h5
Ippl{0}> total number of particles = 999
Ippl{0}> * Restart Energy = 78.4528 (MeV), Path lenght = 37.8059 (m)
Ippl{0}> Tracking Step since last bunch injection is 0
Ippl{0}> 1 Bunches(bins) exist in this file
Ippl{0}> * Gamma = 1.08361, Beta = 0.385189
Ippl{0}> Restart from hdf5 file generated by OPAL-cycl

Do you have a clue?

You could use h5dump to check how many particles are stored in the h5 file.

Every dataset only contains 999 particles (also step 0). Thus, reading is correct. But I checked the writing. The particle with ID 0 is skipped. So, everything's fine. Thanks for the hint.

New reference committed 60bbcc22.