Coupling ordering issue?

[valassi]

Followup from today's discussion. While debugging #826 Olivier observed some coupling ordering issues
#826 (comment)

This might be the root cause of the zero cross section issue #826 (or maybe not)

Note: in any case this is unrelated to channel-iconfig mapping errors discussed in #826 and #852 and probably responsible for #845 and #856

[oliviermattelaer]

@roiser
Did you check the real/imaginary ordering?

If yes then we can close this issue since the PR is in progress

[roiser]

Hi, I can confirm that its all correct for the independent couplings. I need to check for the running ones, cheers

[roiser]

Hi @oliviermattelaer, for after the vacation ;-), I checked the ordering of all the couplings and also the real/imaginary parts, this looks all correct. E.g. looking at the parameters as provided by Fortran (log file) those are

GC_10 =  -0.12781E+01   0.00000E+00
GC_11 =   0.00000E+00   0.12781E+01
GC_13 =   0.00000E+00   0.57824E-04
GC_40 =  -0.00000E+00  -0.28804E+00
GC_53 =  -0.00000E+00  -0.27437E-01
GC_74 =  -0.00000E+00  -0.17058E-01

checking their values as they arrive for the wavefunctions, those are

(gdb) print COUPs
$28 = {0x53ee580, 0x53ee5c0, 0x53ee600, 0x7ffff7f9adc0 <mg5amcCpu::cIPC>, 0x7ffff7f9add0 <mg5amcCpu::cIPC+16>, 0x7ffff7f9ade0 <mg5amcCpu::cIPC+32>}
(gdb) print *COUPs[0]@8 // --> GC_10
$30 = {-1.2458026581298607, -1.2458056165991431, -1.2458036092599556, -1.2458080782619843, 0, 0, 0, 0}
(gdb) print *COUPs[1]@8 // --> GC_11
$31 = {0, 0, 0, 0, 1.2458026581298607, 1.2458056165991431, 1.2458036092599556, 1.2458080782619843}
(gdb) print *COUPs[2]@8 // --> GC_13
$32 = {0, 0, 0, 0, 5.4935752179696128e-05, 5.493601309730976e-05, 5.4935836063073307e-05, 5.4936230200321482e-05}
(gdb) print *COUPs[3]@2 // --> GC_40
$34 = {-0, -0.28804415396362731}
(gdb) print *COUPs[4]@2 // --> GC_53
$35 = {-0, -0.027436627757416142}
(gdb) print *COUPs[5]@2 // --> GC_74
$36 = {-0, -0.017058022507245231}

NB: I'm not sure about the values in the log file above for GC_11/12/13, but I followed the calculation of those with the Gs in the debugger and this looks all correct, and especially about your question the real and imaginary parts of those are in the right order.

NB2: for checking this situation I created a process where madgraph gives me a mix of dependent and independent couplings as they appear in the model, i.e.

generate g g > h > b b~
add process q q > Z Z
add process g g > t t~ 

cheers Stefan

[valassi]

The specific issue 826 about a zero cross section in susy process is fixed by PR #918 (thanks @oliviermattelaer @roiser). Indeed it is caused by a couplings order #862, i.e. this bug.

Closing this as fixed by PR #918 (to be merged soon). Code regenerated in #934.

[valassi]

Attached to #918, reopened and closed again