Automagic integration produces wrong internal analysis results

[dwiesler]

In the attached sindarin example, only the version where automagic integration of an unstable particle is activated produces the correct results. If the integration is requested manually, the event generation takes about a factor of 100 longer and the internally produced histogram is obviously wrong.

I'm not sure wether this is a specific 'feature' of my PSSSM version, I'm trying to reproduce this with an edited version of the casc_dec.sin example.

.sin - example:

model = PSSSM

process dec_lqino_tb    = lqino1 => u,e1,neu1 { $restrictions = "3+4~'se1-'" }
process sig1 = u,g => SE11, lqino1

compile

?slha_read_decays = true
read_slha("sps1aNMSSMed.in")

!if integrate is not commented out, the histogram gets faulty and event generation takes longer by a factor of 100?!
!integrate (dec_lqino_tb)

 mlqino_1 = 600
 wlqino_1 = 5.87
 yuk_nd = 0.312
 yuk_d1 = 0.312
 yuk_d2 = 0
 yuk_dc = 0.312

sqrts = 14000
beams = p, p => lhapdf

integrate (sig1) { iterations = 5:50000, 2:50000 }

n_events = 10000

histogram inv_mass1_tb (0,1000,10)

unstable lqino1 (dec_lqino_tb)

simulate (sig1) {

$sample = "test"
sample_format = lhef
checkpoint = 1000

analysis =
  record inv_mass1_tb (eval M [combine[u,e1]])
}

write_analysis
compile_analysis { $out_file = "sigtehest.dat"}

[dwiesler]

Stupid stupid me, the faulty histograms were due to the mass parameter set AFTER the first (manual) integration, confirming that in my case the problem is mostly located between the keyboard and the chair...

The speed issues seem to depend on the amount of parameters defined after the manually requested decay integration. This is strange, but not relevant for the output, thus I close this ticket for good!