Prepare test scripts for "launch" (and, prior to that, document "launch" internals)

[valassi]

Prepare test scripts for "launch" (and, prior to that, document "launch" internals)

This is a preliminary step to testing "launch" in #659. First we should prepare test scripts for that (and even before, personally I need to understand - reverse engineer - which files are called in which order).

Also related to Stefan's WIP MR #620

[valassi]

Note, I have extended/replace MR #620 with WIP #688, I will try to start from there

[valassi]

Advancing in steps. If in ggttmad (as of MR #688) I try generate events, I get a crash

[avalassi@itscrd80 gcc11.2/cvmfs] /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad> ./bin/generate_events 
...
Error when reading /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/G1/results.dat
Command "generate_events " interrupted with error:
Exception : Reported error: End code 134.0 
         Full associated log: 
        (ompnumthreadsNotSetMeansOneThread) DEBUG: OMP_NUM_THREADS is not set: will use only 1 thread
        (ompnumthreadsNotSetMeansOneThread) omp_get_max_threads() = 1
        ERROR! Card file '../../Cards/param_card.dat' does not exist and environment variable 'MG5AMC_CARD_PATH' is not set
        terminate called after throwing an instance of 'char const*'

[valassi]

Try this

cd ..; rm -rf gg_tt.mad; git checkout gg_tt.mad; cd gg_tt.mad; git status
MG5AMC_CARD_PATH=$(pwd)/Cards ./bin/generate_events 

This is better, but

Command "generate_events " interrupted with error:
FileNotFoundError : [Errno 2] No such file or directory: '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/HTML/run_01'
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/run_01_tag_1_debug.log'.
Please attach this file to your report.

[valassi]

I have added HTML/.keepme. Now the same commands above succeed! But I have no idea what they do.

Note the following changes in the repo at this stage (maybe there's more):

Changes not staged for commit:
  (use "git add/rm <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
        deleted:    Cards/delphes_trigger.dat
        modified:   Cards/param_card.dat
        deleted:    Cards/plot_card.dat
        modified:   Source/make_opts
        modified:   Source/param_card.inc
        modified:   SubProcesses/proc_characteristics
        modified:   SubProcesses/randinit

Untracked files:
  (use "git add <file>..." to include in what will be committed)
        Cards/.delphes_trigger.dat
        Cards/.plot_card.dat
        Events/run_01/
        HTML/results.pkl
        HTML/run_01/
        SubProcesses/P1_gg_ttx/G1/
        SubProcesses/P1_gg_ttx/G1a0/
        SubProcesses/P1_gg_ttx/G1b0/
        SubProcesses/P1_gg_ttx/G2/
        SubProcesses/P1_gg_ttx/G2a0/
        SubProcesses/P1_gg_ttx/G2b0/
        SubProcesses/P1_gg_ttx/G2c0/
        SubProcesses/P1_gg_ttx/G2d0/
        SubProcesses/P1_gg_ttx/G2e0/
        SubProcesses/P1_gg_ttx/G2f0/
        SubProcesses/P1_gg_ttx/G2g0/
        SubProcesses/P1_gg_ttx/G2h0/
        SubProcesses/P1_gg_ttx/G2i0/
        SubProcesses/P1_gg_ttx/G2j0/
        SubProcesses/P1_gg_ttx/G2k0/
        SubProcesses/P1_gg_ttx/ajob1
        SubProcesses/P1_gg_ttx/ajob10
        SubProcesses/P1_gg_ttx/ajob11
        SubProcesses/P1_gg_ttx/ajob12
        SubProcesses/P1_gg_ttx/ajob13
        SubProcesses/P1_gg_ttx/ajob2
        SubProcesses/P1_gg_ttx/ajob3
        SubProcesses/P1_gg_ttx/ajob4
        SubProcesses/P1_gg_ttx/ajob5
        SubProcesses/P1_gg_ttx/ajob6
        SubProcesses/P1_gg_ttx/ajob7
        SubProcesses/P1_gg_ttx/ajob8
        SubProcesses/P1_gg_ttx/ajob9
        SubProcesses/P1_gg_ttx/gensym
        SubProcesses/P1_gg_ttx/input_app.txt
        SubProcesses/P1_gg_ttx/symfact.dat
        crossx.html

One particularly worrying case is param.dat?!

git diff /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/Cards/param_card.dat
--> For patch diff: git diff --no-ext-diff [--no-color]
--> For tkdiff: git difftool
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
epochX/cudacpp/gg_tt.mad/Cards/param_card.dat /tmp/kcTNrF_param_card.dat caf4a67ea8c45bd4d5dc3dbafa15197b3bd7b5f2 100644 epochX/cudacpp/gg_tt.mad/Cards/param_card.dat 0000000000000000000000000000000000000000 100644
2c2
< ## PARAM_CARD AUTOMATICALY GENERATED BY MG5 FOLLOWING UFO MODEL   ####
---
> ## PARAM_CARD AUTOMATICALY GENERATED BY MG5                       ####
4,10d3
< ##                                                                  ##
< ##  Width set on Auto will be computed following the information    ##
< ##        present in the decay.py files of the model.               ##
< ##        See  arXiv:1402.1178 for more details.                    ##
< ##                                                                  ##
< ######################################################################
< 

Looks like param_card.dat is being overwritten (by whom? why?). Clearly related to #660

[valassi]

But this one is also nicely worrying... why is make_opts being overwritten?

[avalassi@itscrd80 gcc11.2/cvmfs] /data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad> git diff --no-ext-diff  Source/make_opts
diff --git a/epochX/cudacpp/gg_tt.mad/Source/make_opts b/epochX/cudacpp/gg_tt.mad/Source/make_opts
index 1c3c559ee..0e7a01dbf 100644
--- a/epochX/cudacpp/gg_tt.mad/Source/make_opts
+++ b/epochX/cudacpp/gg_tt.mad/Source/make_opts
@@ -1,12 +1,17 @@
-GLOBAL_FLAG=-O3 -ffast-math -fbounds-check
+pdlabel1=
+pdlabel2=
+lhapdf=
+PYTHIA8_PATH=NotInstalled
+MG5AMC_VERSION=3.5.0_lo_vect
+GLOBAL_FLAG=-O
+ALOHA_FLAG=
+MATRIX_FLAG=
 DEFAULT_CPP_COMPILER=g++
 MACFLAG=
 STDLIB=-lstdc++
 STDLIB_FLAG=
 DEFAULT_F2PY_COMPILER=f2py
 DEFAULT_F_COMPILER=gfortran
-MG5AMC_VERSION=SpecifiedByMG5aMCAtRunTime
-PYTHIA8_PATH=NotInstalled
 #end_of_make_opts_variables
 
 BIASLIBDIR=../../../lib/

[valassi]

Note: currently generate_event does not return an error code on failure #689.
So adding bash handling that depends on error code will not give the expected results.